454

Second- and third-trimester complications in antiphos-pholipid syndrome (APS) pregnancies are characterizedby placental insufficiency that leads to intrauterinegrowth restriction (IUGR), which, if undetected, may re-sult in fetal death. APS is associated with recurrent mis-carriages and late fetal deaths, and a strong correlationexists between APS and preeclampsia, abruptio placen-tae, and premature delivery. Preeclampsia is often severeand of early onset, and iatrogenic premature deliverymay be indicated because of maternal or fetal complica-tions.1 In the normal pregnancy population, the risk ofthromboembolism is approximately 1 in 1000, whereas

the incidence in APS pregnancies with antithrombotictherapy has been estimated at 4% to 5%.1 Also, there maybe worsening of a preexisting thrombocytopenia. Treat-ment with low-dose aspirin and/or low-molecular-weightheparin (LMWH), close obstetric surveillance, and timelyintervention has improved fetal survival among womenwith APS.1 However, complications still occur frequently.

Much research in the past has focused on endothelialcell activation as a mechanism that contributes to throm-botic events in APS.2,3 Supportive evidence is largely fromin vitro studies, in which cultured human umbilical veinendothelial cells have been exposed to plasma from af-fected patients or antiphospholipid antibodies,3 whichleads to up-regulation of markers of endothelial cell acti-vation. Hitherto no longitudinal investigation of markersof endothelial cell activation in the blood of pregnantwomen with APS has been undertaken, although in asmall cross-sectional study we reported increased serumconcentrations of the cell adhesion molecules soluble in-tercellular adhesion molecule-1 (sICAM-1) and solublevascular cell adhesion molecule (sVCAM-1) in womenwith APS that is uncomplicated by systemic lupus erythe-matosus (SLE),4 when measured at delivery. We now re-port a prospective longitudinal study of a range ofmarkers of endothelial cell activation (Table I) in preg-

From the Departments of Maternal and Fetal Research Unit Women’sHealth, Guy’s, King’s and St Thomas’ School of Medicine, King’s Col-lege,a and the Department of Haematologyb and Lupus PregnancyClinic,c Guy’s and St Thomas’ Hospitals Trust.Supported by Tommy’s the Baby Charity, Lupus UK, and Lifeblood, theThrombosis Charity.Received for publication May 6, 2002; revised June 25, 2002; acceptedSeptember 13, 2002.Reprint requests: L. Poston, PhD, Maternal and Fetal Research Unit,10th Floor North Wing, St Thomas’ Hospital, Lambeth Palace Road,London SE1 7EH, United Kingdom. E-mail: lucilla.poston@kcl.ac.uk© 2003, Mosby, Inc. All rights reserved.0002-9378/2003 $30.00 + 0doi:10.1067/mob.2003.14

Longitudinal evaluation of markers of endothelial celldysfunction and hemostasis in treated antiphospholipidsyndrome and in healthy pregnancy

Sophia Stone, MBBS,a,c Beverley J. Hunt, MD,b,c Paul T. Seed, MSc CStat,a

Kiran Parmar, AIBMS,b Munther A. Khamashta, MD,c and Lucilla Poston, PhDa

London, United Kingdom

OBJECTIVE: The purpose of this study was to determine whether primary antiphospholipid syndrome preg-nancies are associated with endothelial cell activation in the maternal circulation.STUDY DESIGN: Markers of endothelial cell activation were measured every 4 weeks during pregnancy inthe blood of 23 women with primary antiphospholipid syndrome and 19 control subjects. All women with an-tiphospholipid syndrome received anticoagulant treatment. Plasma concentrations of plasminogen activatorinhibitor type 1, tissue plasminogen activator, soluble intercellular adhesion molecule-1, vascular cell adhe-sion molecule-1, E-selectin, and soluble thrombomodulin were determined by enzyme-linked immunoassay.Concentrations of prothrombin fragments 1+2 and D-dimers were also determined.RESULTS: Three antiphospholipid syndrome pregnancies (13%) were complicated by intrauterine growth re-striction and preeclampsia; one antiphospholipid syndrome pregnancy (4%) was complicated by preterm rup-ture of membranes. Six women with antiphospholipid syndrome (26%) had thrombotic events. Differences inconcentrations of endothelial cell activation markers between antiphospholipid syndrome and control preg-nancies were not significant.CONCLUSION: Despite poorer pregnancy outcome, there was no evidence of greater endothelial cell activa-tion in antiphospholipid syndrome pregnancies that were treated. (Am J Obstet Gynecol 2003;188:454-60.)

Key words: Anticardiolipin antibody, lupus anticoagulant, endothelial cell activation, thrombosis,pregnancy complication

Volume 188, Number 2 Stone et al 455Am J Obstet Gynecol

nant women with a confirmed diagnosis of primary APS,as classified by recently agreed on international criteria.5

Concentrations of plasminogen activator inhibitor type 1(PAI-1), tissue plasminogen activator (t-PA; Biopool In-ternational, Ventura, Calif), sICAM-1, sVCAM-1, E-se-lectin (R&D Systems), and soluble thrombomodulin(STM; Diagnostica Stago, Axis Shield, Dundee, UK) weredetermined by enzyme-linked immunosorbent assay(ELISA) longitudinally during pregnancy until deliveryand compared with the same indices that were measuredsimultaneously in a cohort of normal pregnant women.In addition, concentrations of prothrombin fragments1+2 (PF1+2; Dade Behring, Marburg, Germany) and D-dimers (Dade Behring) were also determined as an esti-mate of intravascular hemostatic activity.

Material and methods

Local Ethics Committee approval was obtained. Forty-four pregnant women with primary APS were recruitedfrom the lupus pregnancy clinic, Guy’s and St Thomas’Hospitals at the first antenatal visit (4-5 weeks of gesta-tion) and compared with a cohort of 30 healthy pregnantwomen who were recruited from a community antenatalclinic (first visit approximately 8-11 weeks of gestation).Eleven women were excluded from analysis subsequentlybecause they did not meet the classification criteria of pri-mary APS on the basis of only one positive antiphospho-lipid antibody test. Twenty-eight women consented toserial blood sampling. Of these, 3 women had miscar-riages, and 2 women provided only one sample and wereexcluded from analysis. Control subjects were screenedfor abnormalities of the uterine artery Doppler waveformat 20 and 24 weeks of gestation using pulsed-waveDoppler imaging. Because the presence of unilateral orbilateral notching and a raised resistance index are pre-dictive of preeclampsia and/or IUGR, any controlwomen with either or both of these abnormalities of theDoppler waveform were not entered into the study.Eleven control subjects were excluded from the study, 3who had persistent notching of the Doppler waveform at24 weeks, 1 who had high titers of immunoglobulin G

(IgG) anticardiolipin antibody antibodies, and 7 whowere lost to follow-up. Twenty-three women with APS and19 control subjects provided blood samples every 4 weekson five to eight occasions throughout pregnancy. Controlwomen were matched for parity and ethnicity with no his-tory of adverse pregnancy events, hypertension, or dia-betes mellitus.

APS was defined by the recent International ConsensusCriteria.5 This encompasses a history of thrombosis (ve-nous, arterial, or small vessel) and/or pregnancy morbid-ity (ie, ≥3 consecutive early miscarriages, ≥1 fetal death at>10 weeks of gestation, or premature delivery [<34 weeksof gestation]) because of severe preeclampsia or placen-tal insufficiency that is supported by laboratory evidenceof the disease (lupus anticoagulant and/or anticardi-olipin antibody of IgG and/or IgM isotype present inmedium or high titers on ≥2 occasions at least 6 weeksapart). Twelve women had a history of 13 thromboticevents (7 venous, 6 cerebrovascular accidents); 16 womenhad previous adverse pregnancy events (6 early recurrentmiscarriages, 9 women with ≥1 late fetal death, 2 womenwith preeclampsia, 2 women with small-for-gestational-age infants). Preeclampsia was defined by the Interna-tional Society for the Study of Hypertension in Pregnancyguidelines,6 and small for gestational age was defined asinfants born at ≤10th percentile for gestation and sex,corrected for maternal height, weight, parity, and ethnic-ity with percentile charts.7

Women with APS were treated according to the Guy’sand St Thomas’ Hospitals Lupus Pregnancy Clinic proto-col. This involved administration of 75 mg of aspirin frompreconception plus LMWH (dalteparin, 5000 IU oncedaily) if there was a history of thrombosis, previous latefetal loss, or recurrent miscarriages, despite aspirin ther-apy. In women with past venous thromboembolic events,dalteparin was increased to 5000 IU twice daily from 20weeks of gestation. In those women with previous cere-bral arterial or microvascular events, dalteparin wasstarted at a dose of 5000 units twice daily. If they had neu-rologic events on this dose, they were switched to war-farin in the second trimester, maintaining an

Table I. Markers of endothelial cell dysfunction and hemostasis measured and their function

Marker Function

sICAM-1 Expressed by endothelium, mediates extravasation of leukocytes during inflammationsVCAM-1 Expressed by endothelium, mediates extravasation of leukocytes during inflammationSoluble E-selectin Mediates tethering of circulating leukocytes to endothelium, expressed by endothelial cells on acute

activation, down-regulated by chronic activationSTM Endothelial surface receptor for thrombin responsible for activating protein C, cleaved from surface when

endothelium is damagedt-PA Main physiologic activator of plasminogen in vasculature, produced by endotheliumPAI-1 Inhibitor of t-PA–mediated proteolytic cascade reseased by endothelial cells and activated plateletsPF1+2 Product of prothrombin degradation, measure of amount of thrombin generated during coagulation

activationD-dimers Degradation product of clotting cascade, marker of intravascular fibrinolysis

456 Stone et al February 2003Am J Obstet Gynecol

international normalized ratio of 2.5, with twice weeklytesting. Women were reviewed monthly, and fetal moni-toring was instigated from 20 weeks with timely interven-tion when complications occurred. All control subjectswere untreated.

Maternal blood from study patients was drawn into Va-cutainer (Becton Dickinson, Oxford, UK) sterile tubeswith the minimal-stasis venipuncture technique. Forplasma, blood was collected into sodium citrate; forserum, blood was collected into a Vacutainer SST tube(Becton Dickinson) containing clot activator and gel.Samples were centrifuged (2215g, 15 minutes, 4°C), andaliquots of plasma and serum were stored at –70°C untilanalysis. Five milliliters of the citrated sample super-natant was centrifuged again to obtain platelet-poorplasma. Anticardiolipin IgG and IgM were detected byELISA. Lupus anticoagulant was detected according tothe criteria defined by Brandt et al.8 With the dilute Rus-sell’s viper venom time, the presence of lupus anticoagu-lant was confirmed if the addition of the phospholipidshortened the prolonged clotting time by >10%.

The following markers of endothelial cell activationwere measured by ELISA: PAI-1 and t-PA (interassay vari-ation, 2.9% and 4.0%, respectively; intra-assay variation,

3.3% and 5.0%, respectively), sICAM-1 and sVCAM-1 (in-terassay variation, 7.4% and 9.2%, respectively; intra-assayvariation, 3.6% and 5.0%, respectively), E-selectin (in-terassay variation, 7.4%; intra-assay variation, 4.9%), andSTM (interassay variation, 4.0%; intra-assay variation,4.8%). Serum samples were used to measure sICAM-1,sVCAM-1, and E-selectin; plasma samples were used forSTM, PAI-1, and t-PA. Because endothelial cell activationevokes thrombotic changes, the markers of intravascularhemostatic turnover, PF1+2 and D-dimers, were also de-termined with the use of plasma (interassay variation,9.0% and 9.4%, respectively; intra-assay variation, 6.0%and 4.0%, respectively).

Patient and control samples were analyzed on the sameplate to minimize plate-to-plate variation. Data were en-tered into a Microsoft Excel (Microsoft Corporation,Redmond, Wash) database.

Statistical analysis. In the absence of any directly rele-vant study, power calculations were calculated with theuse of data that were obtained in a previous study fromour unit of sICAM-1 and sVCAM-1.4 On the basis of thedifferences observed between women with APS and con-trol subjects for sVCAM-1, which require the greaternumber of subjects to achieve adequate power, we aimed

Table II. Clinical history and pregnancy outcome of women with primary APS

APS Current pregnancy

APL, LA/ Gestational age Birth weight (g) No. aCL History Treatment Complications at delivery (wk) (percentile)

1 LA IUGR, FD �2 Aspirin IUGR, PE 33.0 1555 (1.2)2 LA, aCL CVA Aspirin, LMWH Thrombotic 37.5 2800 (31.6)

microangiopathy3 LA MISC �8 Aspirin, LMWH, PPROM at 34 wk 35.9 2410 (25.1)

prednisolone 10 mg to 14 wk

4 LA CVA Warfarin TIA, switch to warfarin 38.2 2840 (4.6)5 aCL VTE �2 Aspirin, LMWH 40.9 3630 (43.7)6 LA, aCL VTEs, CVA, FD �5 Warfarin TIA when noncompliant 39.4 3140 (11.6)7 LA IUGR, FD Aspirin, LMWH 38.1 3080 (62.4)8 aCL FD, VTE Aspirin, LMWH 38.7 2640 (2.0)9 aCL MISC �4 Aspirin, LMWH Hypertension 38.6 3650 (66.9)

10 aCL MISC �4 Aspirin, LMWH 39.4 4649 (100)11 LA MISC �5 Aspirin, LMWH 42.0 3400 (20.9)12 LA, aCL VTE Aspirin, LMWH TIA at 38 wk 38.0 2500 (3.6)13 LA, aCL FD Aspirin, LMWH PE, IUGR, abruptio placentae 34.3 1900 (2.7)14 LA VTE Aspirin, LMWH TIA when noncompliant 40.1 3020 (10.9)15 aCL VTE Aspirin, LMWH 40.3 4640 (98.4)16 β2GPI PE, FD Aspirin, LMWH Hypertension 36.0 2665 (42.4)17 LA MISC �4 Aspirin, LMWH 39.0 2610 (6.6)18 LA, aCL CVA, PE, IUGR Aspirin, LMWH TIA, PE, IUGR 33.0 1673 (4.3)19 aCL CVA Warfarin TIA, switch to warfarin 38.3 2850 (15.8)20 LA MISC �4 Aspirin, LMWH 40.0 3544 (43.5)21 aCL FD �2 Aspirin, LMWH 38.3 3500 (49.6)22 aCL FD �2 Aspirin, LMWH 38.4 3800 (91.4)23 LA VTE, FD Aspirin, LMWH Proteinuria 38.4 2720 (24.4)

APS, n = 23; mean (±SD) gestational age at delivery (wk) 38.1 (±2.3), mean (±SD) birth weight (g) 3009 (±793), mean (±SD) calculatedpercentile 33.2 (±31.9). Control, n = 19; mean (±SD) gestational age at delivery (wk) 40.4 (±0.7) (P = .002), mean (±SD) birth weight (g)3496 (±533) (P = .008), mean (±SD) calculated percentile 48.0 (±36.3) (P = .098). LA, Lupus anticoagulant; aCL, anticardiolipin anti-body; IUGR, intrauterine growth restriction; FD, fetal death; PE, preeclampsia; CVA, cerebrovascular accident; MISC, miscarriage;PPROM, preterm premature rupture of membranes; TIA, transient ischemic attack; VTE, venous thromboembolism; B2GPI, B2-glyco-protein I.

Volume 188, Number 2 Stone et al 457Am J Obstet Gynecol

to recruit 30 subjects with APS in this study to achieve90% power to detect a 30% difference in the indices thatwere evaluated with a probability of <.05.

Data were transferred to Stata (version 7.0; Stata Corp,College Station, Tex) for analysis. All blood markers

were found to have approximately log-normal distribu-tions. Analysis was conducted on the logged values, andresults are presented as geometric means with CIs. Thedata were grouped in monthly gestational intervals from3 months to term, to allow an analysis of repeated mea-

Fig 1. Gestational profile of cell adhesion molecules. A, sICAM-1in women with APS (open circles, solid line) compared with controlsubjects (open squares, dashed line; sICAM-1, –6%; 95% CI, –22%to +11%; P = .462). B, sVCAM-1 in women with APS (open circles,solid line) compared with control subjects (open squares, dashedline; sVCAM-1, +5%; 95% CI, –9% to +21%; P = .492). C, SolubleE-selectin in women with APS (open circles, solid line) and controlsubjects (open squares, dashed line); APS compared with control (E-selectin, –1%; 95% CI, –27% to +32%; P = .913). Data are pre-sented on log scale and given as geometric means, with bars thatare based on 1 SD. Probability values refer to comparison ofgroups over entire period of gestation that was studied.

Fig 2. Gestational profile of markers of endothelial cell activationand fibrinolytic activity. A, STM in women with APS (open circles,solid line) compared with control subjects (open squares, dashedline; STM, –7%; 95% CI, –18% to +8%; P = .385). B, PAI-1 inwomen with APS (open circles, solid line) compared with controlsubjects (open squares, dashed line; PAI-1, +4%; 95% CI, –13% to+38%; P = .451). C, t-PA in women with APS (open circles, solid line)and control subjects (open squares, dashed line) with APS com-pared with control subjects (t-PA, +25%; 95% CI, –1% to +65%;P = .058). Data are presented on log scale and given as geomet-ric means, with bars based on 1 SD. Probability values refer tocomparison of groups over entire period of gestation that was

458 Stone et al February 2003Am J Obstet Gynecol

sures with the use of generalized estimating equations9

with robust SEs and correction for any increase or de-crease that was observed in the pregnant control sub-jects. The ratio of the geometric means (multiples of themedian) between APS and control subjects was assessed.Statistical significance was assumed when a probabilityvalue of ≤.05.

Results

Pregnancy outcome. Clinical data and pregnancy out-come are shown in Table II. All women with APS (meanage, 33.9 ± [SD] 4.3 years) received aspirin and LMWH.Three of the five women who had transient ischemic at-tacks were treated with warfarin from the secondtrimester. One woman had postpartum thrombotic mi-croangiopathy of the kidney. All control pregnancies(mean age, 29.4 ± 5.3 years) were uncomplicated.

The mean gestation at delivery (P < .005) and meanbirth weight (P = .008) were lower in the APS group.

Labor was induced in 39% of women with APS, but innone of the control subjects. The caesarean delivery ratewas 48% (11/23 deliveries, 7 emergency) in women withAPS and 25% in control pregnancies (5/20 deliveries, 4emergency). Although the livebirth rate was 100% in theAPS group, four infants (17%) were admitted to theneonatal care unit.

Markers of endothelial cell activation. The concentra-tions of the endothelial cell activation markers are shownin Figs 1 through 3. Differences in concentrations of themarkers between the APS and control samples were allsmall and nonsignificant, although there was a trend to-wards a higher concentration of t-PA in the APS group (P = .058). The concentrations of PAI-1, t-PA, PF1+2, D-dimers, and STM increased with gestational age in bothgroups. When they were subanalyzed according to treat-ment (aspirin alone vs aspirin and heparin vs warfarin),D-dimer concentrations were lower in the women who re-ceived warfarin (P = .043); when analyzed according topregnancy complications (maternal [ie, thromboticevents] and fetal [including IUGR and preeclampsia]against normal outcome in the APS group), only PF1+2concentrations differed significantly in the women whohad thrombotic events (P = .018). However, patient num-bers in all subanalysis groups were small.

Comment

Endothelial cell activation is a pathologic factor thatcontributes to a number of vascular diseases and pro-thrombotic conditions, including preeclampsia, one ofthe most common complications of APS pregnancies. En-dothelial cell activation also occurs in patients with SLE,10

of whom 25% have antiphospholipid antibodies. In thisstudy, we hypothesized that abnormal pregnancy out-come in women with APS may be associated with en-dothelial cell activation. Despite treatment, nine womenhad complications, and the mean birth weight and gesta-tional age at delivery were significantly lower than thoseof the control women. Nonetheless, the data presentedhere strongly suggest that the activation of the endothe-lial cell layer is not a major contributory factor to the ab-normal pregnancy outcome in these women with APSwho were treated. The nonsignificant trend in concentra-tions of t-PA between the study group and the controlgroup may reflect a type II error, and it is possible thatlarger study numbers might reveal a more pronounceddifference. However, the lack of any generalized defect inendothelial cell activation would concur with another re-cent study from our laboratory that has shown no differ-ence in levels of sVCAM-1, sICAM-1, interleukin-6,endothelin-1, von Willebrand Factor, and P-selectin in 59nonpregnant patients with thrombosis and primary APSwho were treated with warfarin when compared withhealthy control subjects,11 which substantiates a previousreport in nine nonpregnant patients with APS (7 of

Fig 3. Gestational profile of markers of intravascular hemostaticactivity. A, PF1+2 in women with APS (open circles, solid line) com-pared with control subjects (open squares, dashed line; PF1+2, –9%;95% CI, –23% to +13%; P = .505). B, D-dimers in women withAPS (open circles, solid line) and control subjects (open squares,dashed line); APS compared with control subjects (D-dimers,+4%; 95% CI, –43% to +121%; P = .735). Data are presented ona log scale and given as geometric means, with bars based on 1SD. Probability values refer to a comparison of groups over en-tire period of gestation that was studied.

Volume 188, Number 2 Stone et al 459Am J Obstet Gynecol

whom were receiving warfarin)12 in which normal con-centrations of t-PA, PAI-1, von Willebrand Factor, andantiendothelial cell antibodies were observed. Frijns etal10 also found no association between endothelial cell ac-tivation and the presence of antiphospholipid antibodies(n = 37) or a history of thrombosis (n = 38) in 76 patientswith SLE, but no mention was made of antithrombotictherapy.

In contrast, a few recent reports have indicated that en-dothelial cell activation is associated with APS in preg-nant and nonpregnant subjects. An earlier study fromour group showed an increase in sVCAM-1 and sICAM-1in samples taken at delivery from 8 treated women withAPS, but gestational age at delivery varied widely, and thesamples therefore were not gestation matched with thecontrol subjects.4 In addition, Martinuzzo et al13 haveshown an elevation of VCAM-1, E-selectin, and P-selectinin 67 nonpregnant patients with antiphospholipid anti-bodies that were uncomplicated by SLE, of whom 20 pa-tients were receiving oral anticoagulant treatment, butthere was no relationship between concentrations of en-dothelial cell activation markers and either a history ofthrombosis or an antibody profile. An elevated concen-tration of sVCAM-1 was also demonstrated in the blood of24 nonpregnant patients with primary APS that was takenat the time of or in the days after a thrombotic event.14

This may have been the result of endothelial injury, how-ever, which resulted from the thrombotic event andtherefore not a primary process in thrombogenesis.

The first suggestion of endothelial cell activation as apotential trigger to thrombosis in APS was the detectionof antiendothelial cell antibodies in a group of nonpreg-nant patients with APS with venous and/or arterialthrombosis15; however, samples were obtained in the firstweek after the occurrence of the thrombotic event. Oth-ers subsequently showed evidence of endothelial cell activation caused by the addition of antiphospholipid an-tibodies in vitro with tumor necrosis factor–induced pro-coagulant activity,16 the up-regulation of cell adhesionmolecule expression in cultured human umbilical venousendothelial cells,3 and the in vitro and in vivo activationof endothelial cells by antiphospholipid antibodies in amouse model.2 However, the expression of endothelialcell activation markers in cultured endothelial cells maydiverge from that in vivo, and the relevance of humanumbilical vein endothelial cells to maternal endotheliumis questionable. Earlier suggestions of endothelial cell ac-tivation in APS may also have been confounded by a lackof accurate classification, standardization of the antiphos-pholipid assays, and the confounding influence of SLE.

It must also be considered that the recent introductionof routine thromboprophylaxis may have contributed tothe discrepancies in earlier and later studies in relation toendothelial cell activation in APS, but there is little evi-dence to support an inhibitory effect of anticoagulatory

drugs on endothelial cell activation. High-dose aspirinthat leads to serum salicylate concentrations of 1 to 5mmol/L inhibits the activation of NF-κB, a transcriptionfactor that is implicated in endothelial cell activation,17

but there is no report of the effects of low-dose aspirin.Only one investigation in dogs, which studied the effectof heparin on endothelial cell activation, suggested thatheparin may reduce the activation of NF-κB18; but no sim-ilar studies exist in man, and none of the studies have in-vestigated the effect of warfarin on endothelial cellactivation. Both Bombeli et al19 and another recent studyof 81 nonpregnant patients who were receiving war-farin20 found that anticoagulant treatment influencedthe concentration of D-dimers. By subanalysis, our studyshowed that D-dimer concentrations were 62% lower inthe women receiving warfarin therapy (95% CI, 3%-85%;P = .043). Although numbers are small, this suggests thatmarkers of hemostatic turnover are unlikely to be helpfulin the prediction of pregnancies that are at risk of throm-botic events in patients who are treated with thrombo-prophylaxis.

The question remains therefore whether potential en-dothelial cell activation, if present in APS, is a primaryprocess that is dependent on the presence of antiphos-pholipid antibodies or a secondary process that is depen-dent on the actual initiation of the thrombotic processes.Thus, in this cohort, the negative findings could havebeen due to treatment that blunted the initiation of a cas-cade of clotting events.

No study has determined longitudinal gestational pro-files simultaneously in normal pregnancy of the endothe-lial cell activation and the hemostatic indices that wereevaluated here, and some interesting patterns haveemerged. Concentrations of all three hemostatic markersincreased with gestational age; D-dimers have previouslybeen shown to increase with gestation in uncomplicatedpregnancy21 and in pregnant women at risk of thrombo-sis,19 whereas PF1+2 has only been evaluated previously innormal pregnancy in a cross-sectional study.22 The en-dothelial cell activation markers t-PA and STM increasedwith gestation and may reflect a degree of maternal en-dothelial cell activation in normal pregnancy. PAI-1 con-centrations were also observed to rise with gestation; butbecause PAI-1 is considered to originate from the utero-placental endothelium, the rise in normal pregnancy ismore likely to reflect an increasing placental mass withadvancing gestation than endothelial cell activation. Wehave observed a similar gestational profile of PAI-1 in nor-mal pregnant women in a previous longitudinal study inwhich we also reported raised concentrations in womenwho later had preeclampsia.23 t-PA21 and STM24 haveeach been measured previously in only 1 longitudinalstudy; a similar gestation-related increase in normal preg-nancy was observed. Krauss et al25 established gestationalplasma profiles of sICAM-1 and sVCAM-1 in a cross-

460 Stone et al February 2003Am J Obstet Gynecol

sectional study; in agreement with our data, these mark-ers showed little association with gestation during normalpregnancy.

In conclusion, this study provides evidence for mild en-dothelial cell activation in normal pregnancy but doesnot support a major role for maternal endothelial cell ac-tivation in the development of maternal and fetal compli-cations in treated APS pregnancies. The investigation hasalso highlighted the improved outcome of APS pregnan-cies since the introduction of standard treatment regimesand close obstetric vigilance in a tertiary referral unit withextensive experience in treating women with APS.

REFERENCES

1. Branch DW, Silver RM, Blackwell JL, Reading JC, Scott JR. Out-come of treated pregnancies in women with antiphospholipidsyndrome: an update of the Utah experience. Obstet Gynecol1992;80:614-20.

2. Pierangeli SS, Colden-Stanfield M, Liu X, Barker JH, AndersonGL, Harris EN. Antiphospholipid antibodies from antiphospho-lipid syndrome patients activate endothelial cells in vitro and invivo. Circulation 1999;99:1997-2002.

3. Simantov R, LaSala JM, Lo SK, Gharavi AE, Sammaritano LR,Salmon JE, et al. Activation of cultured vascular endothelial cellsby antiphospholipid antibodies. J Clin Invest 1995;96:2211-9.

4. Lakasing L, Campa JS, Parmar K, Poston R, Hunt BJ, Poston L.Normal expression of cell adhesion molecules in placentae fromwomen with systemic lupus erythematosus and the antiphospho-lipid syndrome. Placenta 2000;21:142-9.

5. Wilson WA, Gharavi AE, Koike T, Lockshin MD, Branch DW,Piette JC, et al. International consensus statement on prelimi-nary classification criteria for definite antiphospholipid syn-drome. Arthritis Rheum 1999;42:1309-11.

6. Davey DA, MacGillivray I. The classification and definition of thehypertensive disorders of pregnancy. Am J Obstet Gynecol 1988;158:892-8.

7. Gardosi J, Chang A, Kalyan B, Sahota D, Symonds EM. Cus-tomised antenatal growth charts. Lancet 1992;339:283-7.

8. Brandt JT, Triplett DA, Alving B, Scharrer I. Criteria for the di-agnosis of lupus anticoagulants: an update. On behalf of theSubcommittee on Lupus Anticoagulant/Antiphospholipid Anti-body of the Scientific and Standardisation Committee of theISTH. Thromb Haemost 1995;74:1185-90.

9. Liang K-Y, Zeger SL. Longitudinal analysis using generalised lin-ear models. Biometrika 1986;73:13-22.

10. Frijns CJ, Derksen RH, de Groot PG, Algra A, Fijnheer R. Lupusanticoagulant and history of thrombosis are not associated withpersistent endothelial cell activation in systemic lupus erythe-matosus. Clin Exp Immunol 2001;125:149-54.

11. Williams FM, Parmar K, Hughes GR, Hunt BJ. Systemic en-dothelial cell markers in primary antiphospholipid syndrome.Thromb Haemost 2000;84:742-6.

12. Mackworth-Young CG, Andreotti F, Harmer I, Loizou S, Pot-tinger BE, Pearson JD, et al. Endothelium-derived haemostaticfactors and the antiphospholipid syndrome. Br J Rheumatol1995;34:201-6.

13. Martinuzzo ME, Forastiero RR, Kordich L, Carreras LO. In-creased lipid peroxidation correlates with platelet activation butnot with markers of endothelial cell and blood coagulation acti-vation in patients with antiphospholipid antibodies. Br J Haema-tol 2001;114:845-51.

14. Kaplanski G, Cacoub P, Farnarier C, Marin V, Grégoire R, GatelA, et al. Increased soluble vascular cell adhesion molecule 1 con-centrations in patients with primary or systemic lupus erythe-matosus-related antiphospholipid syndrome. Arthritis Rheum2000;43:55-64.

15. Cervera R, Khamashta MA, Font J, Ramirez G, D’Cruz D, Mon-talban J, et al. Antiendothelial cell antibodies in patients with theantiphospholipid syndrome. Autoimmunity 1991;11:1-6.

16. Oosting JD, Derksen RHWM, Blokzijl L, Sixma JJ, de Groot PG.Antiphospholipid antibody positive sera enhance endothelialcell procoagulant activity: studies in a thrombosis model.Thromb Haemost 1992;68:278-84.

17. Yin MJ, Yamamoto Y, Gaynor RB. The anti-inflammatory agentsaspirin and salicylate inhibit the activity of I(kappa)B kinase-beta. Nature 1998;396:77-80.

18. Thourani VH, Brar SS, Kennedy TP, Thornton LR, Watts JA,Ronson RS, et al. Nonanticoagulant heparin inhibits NF-kappaBactivation and attenuates myocardial reperfusion injury. Am JPhysiol Heart Circ Physiol 2000;278:H2084-93.

19. Bombeli T, Raddatz-Mueller P, Fehr J. Coagulation activationmarkers do not correlate with the clinical risk of thrombosis inpregnant women. Am J Obstet Gynecol 2001;184:382-9.

20. Ombandza-Moussa E, Samama MM, Horellou MH, ChatelierAL, Elalamy I, Conard J. Influence of oral anticoagulanttreatment on D-dimers levels. Ann Biol Clin (Paris) 2001;59:579-83.

21. Bellart J, Gilabert R, Fontcuberta J, Borrell M, Miralles RM,Cabero L. Fibrinolysis changes in normal pregnancy. J PerinatMed 1997;25:368-72.

22. Cadroy Y, Grandjean H, Pichon J, Desprats R, Berrebi A,Fournie A, et al. Evaluation of six markers of haemostatic sys-tem in normal pregnancy and pregnancy complicated by hy-pertension or pre-eclampsia. Br J Obstet Gynaecol 1993;100:416-20.

23. Chappell L, Seed PT, Briley AL, Kelly FJ, Hunt BJ, Charnock-Jones DS, et al. A longitudinal study of biochemical variables inwomen at risk of pre-eclampsia. Am J Obstet Gynecol 2002;187:127-36.

24. de Moerloose P, Mermillod N, Amiral J, Reber G. Thrombo-modulin levels during normal pregnancy, at delivery and in thepostpartum: comparison with tissue-type plasminogen activatorand plasminogen activator inhibitor-1. Thromb Haemost1998;79:554-6.

25. Krauss T, Kuhn W, Lakoma C, Augustin HG. Circulating en-dothelial cell adhesion molecules as diagnostic markers for theearly identification of pregnant women at risk for developmentof preeclampsia. Am J Obstet Gynecol 1997;177:443-9.