Evidence of maternal platelet activation, excessive thrombin generation, and high amniotic fluid tissue factor immunoreactivity and functional activity in patients with fetal death

Abstract

Objective. Fetal death can lead to disseminated intravascular coagulation or fetal death syndrome. However, currently it is not clear what are the changes in the coagulation system in patients with a fetal death without the fetal death syndrome. This study was undertaken to determine: (1) whether fetal death in the absence of fetal death syndrome is associated with changes in hemostatic markers in maternal plasma and amniotic fluid; and (2) whether maternal hypertension or placental abruption are associated with further changes in the hemostatic profile of these patients.

Methods. A cross-sectional study included the following: (1) determination of changes in markers of coagulation and platelet activation in patients with a normal pregnancy (n = 71) and patients with fetal demise (FD) without disseminated intravascular coagulation (n = 65); (2) determination of the amniotic fluid (AF)–tissue factor concentration and activity, as well as the concentrations of thrombin–antithrombin III (TAT) complexes in patients with a normal pregnancy (n = 25) and those with a FD (n = 36) who underwent amniocentesis. Plasma and AF concentrations of TAT complexes and TF (an index of thrombin generation), as well as maternal plasma concentrations of sCD40L (a marker of platelet activation), tissue factor pathway inhibitor (TFPI) and prothrombin fragments (PF) 1 + 2 (also an indicator of in vivo thrombin generation) were measured by ELISA. TF and TFPI activity were measured using chromogenic assays.

Results. (1) patients with FD without hypertension had a higher median maternal plasma sCD40L concentration than normal pregnant women (P < 0.001); (2) patients with FD had a higher median maternal plasma TAT III complexes than women with a normal pregnancy (P < 0.001); (3) the median AF–TF concentration and activity were higher in the FD group than in the normal pregnancy group (P < 0.001 for both); (4) patients with preeclampsia and FD had a higher median maternal plasma immunoreactive TF concentration than both normotensive patients with FD and women with normal pregnancies (P < 0.001 and P = 0.001, respectively); (5) the median plasma TF activity was higher in patients with preeclampsia and FD than that of women with normal pregnancies (P = 0.003); (6) among patients with a FD, those with placental abruption had a higher median AF–TAT complexes concentration than those without abruption (P = 0.0004).

Conclusions. Our findings indicate that: (1) mothers with a FD have evidence of increased in vivo thrombin generation and platelet activation than women with normal pregnancies; (2) patients with a FD and hypertension had a higher degree of TF activation than those with fetal death but without hypertension; (3) the AF of women with a FD had a higher median TF concentration and activity than that of normal pregnant women. AF can be a potential source for tissue factor and it participates in the development of fetal death syndrome in patients with a retained dead fetus.

Keywords:

• Soluble CD40L
• thrombin–antithrombin complexes
• prothrombin fragments 1 + 2
• pregnancy
• abruption
• preeclampsia

Introduction

The term fetal demise syndrome refers to the association between fetal death and maternal coagulopathy [1] reported in the 1950's among patients with Rh isoimmunization [1-6]. Subsequent observations demonstrated that maternal disseminated intravascular coagulation (DIC) develops only if the dead fetus is retained in utero for more than 5 weeks, suggesting a chronic rather than an acute process as the underlying mechanism of disease [4-6]. However, the precise genesis of this chronic process of the fetal death syndrome remains unclear, largely because no adequate assays for tissue factor were available when the pathophysiology of fetal death syndrome was attributed to an excess of tissue thromboplastin generated from the dead fetus.

The changes in the coagulation system during pregnancy are considered to be adaptive mechanisms for the prevention of bleeding at the time of delivery [7-11]. Indeed, normal pregnancy is associated with a substantial increase in TF concentrations in both the decidua and myometrium [12-15]. Similarly, high TF concentrations have been detected in the fetal membranes (mainly the amnion) and amniotic fluid (AF) [7],[16-19]. In contrast to the changes detected in AF and decidua, the maternal plasma immunoreactive TF concentrations of women with a normal pregnancy do not differ significantly from that of non-pregnant women [11],[20]. However, women in labor at term have significantly higher maternal plasma immunoreactive TF concentrations than that of non-pregnant women [19]. In addition to the changes in TF, normal pregnancy is associated with excessive thrombin generation [11],[21], as determined by increased maternal concentrations of fibrinopeptide A, prothrombin fragments (PF) 1 and 2, and thrombin–antithrombin (TAT) III complexes [7],[22-24]. The concentration of these complexes increases further during and after normal labor [25] and delivery [23],[25], and decreases later during the puerperium [23],[25]. Increased thrombin generation has been implicated as a mechanism of disease in several obstetrical syndromes including: preeclampsia [26-33], small-for-gestational age (SGA) [30],[34],[35], preterm labor [21],[36], and preterm prelabor rupture of membranes (PROM) [21],[37].

This study was undertaken to determine whether: (1) fetal death prior to the development of fetal death syndrome is associated with changes in hemostatic markers in the maternal plasma as well as AF; and (2) whether maternal hypertension or placental abruption are associated with further changes in the hemostatic profile of these patients.

The following analytes were studied in the maternal plasma: (1) TF concentration and activity, as well as markers of platelet activation (sCD40L), all needed for thrombin generation; (2) markers of thrombin generation (TAT III complexes, PF 1 + 2); and (3) the concentration and activity of tissue factor pathway inhibitor (TFPI), the extrinsic pathway inhibitor. The intra-amniotic concentrations of TAT III complexes, TF concentrations, and activity were studied as well.

Material and methods

Study groups and inclusion criteria

A cross-sectional study was conducted in two steps: (1) determination of the maternal plasma concentration of TF, TFPI, PF 1 + 2, TAT III complexes and sCD40L, as well as the activity of TF and TFPI in patients with a normal pregnancy (n = 71) and patients with a fetal demise (FD) without (DIC) (n = 65); (2) determination of the AF–TF concentration and activity as well as the TAT complexes concentrations in patients with normal pregnancies and those with a FD. AF was available for analysis from 36 (55.4%) patients in the FD group and 3 (4.2%) patients in the normal pregnancy group. Therefore, we identified 22 patients with a normal pregnancy outcome who underwent amniocentesis at a similar gestational age as patients with a FD, and included them as normal controls.

Samples and data were retrieved from the Perinatology Research Branch bank of biological samples and clinical databases. Many of these samples have been previously employed to study the biology of inflammation, hemostasis, angiogenesis regulation, and growth factor concentrations in non-pregnant women, normal pregnant women, and those with pregnancy complications. All participating women provided a written informed consent prior to the collection of maternal blood and AF. The Institutional Review Boards of both Wayne State University and the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD/NIH/DHHS) approved the collection and utilization of samples for research purposes.

Clinical definitions

Women with normal pregnancies met the following criteria: (1) no medical, obstetrical, or surgical complications at the time of the study; (2) gestational age at blood collection or amniocentesis ranging from 20 to 41 weeks; and (3) delivery of a term neonate, appropriate for gestational age, without complications. Patients with multiple pregnancies or fetuses with congenital and/or chromosomal anomalies were excluded. Fetal death was defined as a FD diagnosed and confirmed by ultrasound ≥ the 19th week of gestation. SGA was diagnosed as a birth weight below the 10th percentile for gestational age [38]. Hypertensive disorders of pregnancy included gestational hypertension and preeclampsia. Gestational hypertension was defined as hypertension (systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg on at least two occasions, 4 h to 1 week apart) that was diagnosed after 20 weeks of gestation without proteinuria. Preeclampsia was defined in the presence of hypertension and proteinuria diagnosed after 20 weeks of gestation (≥300 mg in a 24-h urine collection or at least one dipstick measurement ≥2+) [39]. DIC was defined according to the DIC score of the scientific and standardization committee on disseminated intravascular coagulation of the international society on thrombosis and hemostasis [40],[41]. The parameters included in this score are: platelet count, fibrin related markers, prothrombin time and fibrinogen level, and a score ≥5 is required for diagnosis. Upon admission to the labor and delivery suite (the time when the maternal blood sample was collected), none of the women included in the FD group had an abnormal value in platelet count (<100,000/ml), prothrombin time (>3 s), or fibrinogen level (<1 g/l or 100 mg/dl).

Placental histopathologic examinations

Placental tissue samples were taken by systematic random sampling, subsequently fixed in 10% neutral buffered formalin overnight and embedded in paraffin. Five micrometer-thick paraffin sections, stained with haematoxylin and eosin, were examined using bright-field light microscopy. Three pathologists blinded to the clinical information performed histopathologic examinations, and the placental lesions were classified according to a diagnostic schema proposed by Redline et al. [42]. For the analysis of the association between the different analytes and specific placental lesions, we grouped them into three categories: (1) lesions associated with intra-amniotic infection and/or inflammation; (2) findings associated with maternal under perfusion; and (3) lesions associated with fetal vascular thrombotic occlusive disease.

Sample collection

Maternal blood collection

Blood samples were collected with a vacutainer into 0.109 M trisodium citrate anticoagulant solution (BD; San Jose, CA), centrifuged at 1300 g for 10 min at 4°C, and stored at −70°C until assayed.

Amniotic fluid collection

Amniocentesis was performed at the discretion of the treating physician. The procedure was done trans-abdominally under ultrasonographic guidance in order to determine the microbiologic state of the amniotic cavity, fetal karyotype, injection of AF dye to rule out rupture of the fetal membranes, or for assessment of fetal lung maturity. AF was transported to the laboratory in a capped plastic syringe and cultured for aerobic and anaerobic bacteria as well as genital mycoplasmas. White blood cell (WBC) count, glucose concentration and Gram stain for microorganisms were performed in AF shortly after collection. Intra-amniotic infection was defined by the presence of a positive AF culture for microorganisms and inflammation by an AF WBC count ≥100 cells/ml. The results of the AF analyses were used for clinical management.

Immuno and chromogenic assays

Human tissue factor immunoassay

TF concentrations were determined by sensitive and specific immunoassays obtained from American Diagnostica (Greenwich, CT), which recognizes TF-apo, TF, and TF-FVII complexes. The assays were conducted according to the manufacturer's recommendations. The calculated coefficient of variation (CV) in our laboratory was 5.3%, and the sensitivity was 10 pg/ml.

Human tissue factor activity assay

TF activity was determined by a chromogenic assay obtained from American Diagnostica (Greenwich, CT). The assays were conducted according to the manufacturer's recommendations. In our laboratory, the calculated intra-assay CV was 3.77%, whereas the inter-assay CV was 6.25%, and the sensitivity of this assay was 0.53 pM.

Human tissue factor pathway inhibitor immunoassay

The concentrations of TFPI in maternal plasma were determined by sensitive and specific immunoassays obtained from American Diagnostica (Greenwich, CT). The TFPI ELISA employs, as the capture antibody, a murine anti-TFPI monoclonal that is directed against the Kunitz-1 domain of the TFPI molecule; therefore, it detects both TFPI-1 and TFPI-2, while measuring the total TFPI plasma concentrations. The assay was conducted according to the manufacturer's recommendations. The calculated CV in our laboratory was 6.6%, and the sensitivity was approximately 10 ng/ml.

Human tissue factor pathway inhibitor activity assay

TFPI activity was determined by a chromogenic assay obtained from American Diagnostica (Greenwich, CT). The assays were conducted according to the manufacturer's recommendations. In our laboratory, the calculated intra-assay CV was 5.51%, whereas the inter-assay CV was 8.74%, and the sensitivity was 0.017 unit/ml.

Human sCD40L immunoassay

Maternal plasma sCD40L concentrations were determined by sensitive and specific immunoassays obtained from R&D Systems (Minneapolis, MN). The assays were conducted according to the manufacturer's recommendations. In our laboratory, the calculated intra-assay CV was 5%, whereas the inter-assay CV was 6.2%, and the sensitivity of the sCD40L assay was 4.2 pg/ml.

Human thrombin–antithrombin III complexes immunoassay

Maternal plasma concentrations of TAT III complexes were determined by a specific immunoassay. The manufacturer (Dade Behring, Marburg, Germany) validated this assays for citrated plasma samples, and we conducted these immunoassays following to their instructions. In our laboratory, the calculated inter-assay CV was 7.67% whereas the intra-assay CVs was 5.98% and the sensitivity was 0.81 μg/L.

Human prothrombin fragments 1 + 2 immunoassay

Maternal plasma concentration of PF 1 + 2 was determined by specific immunoassay. The manufacturer (Dade Behring, Marburg, Germany) validated this assay for citrated plasma samples and we conducted these immunoassays according to their instructions. In our laboratory, the calculated inter-assay CV in our laboratory was 5.35%, whereas the intra-assay CVs was 5.98% and the sensitivity was for PF1 + 2 assay 5.33 pmol/L.

Statistical analysis

Tissue factor concentrations and activity, TFPI concentration and activity, sCD40L, TAT III complexes and PF 1 + 2 concentrations in the plasma and/or AF were not normally distributed; thus, Kruskal–Wallis test with post-hoc Mann–Whitney U test were used for comparisons of continuous variables. Comparisons of proportions were performed by Chi-square and Fisher's exact tests. A P value < 0.05 was considered statistically significant. Analyses were performed with SPSS, version 12 (SPSS, Chicago, IL).

Results

Demographic and clinical characteristics of the study groups

Table I displays the demographic and clinical characteristics of the study groups. Patients with a FD had a lower median gestational age at delivery and neonatal birthweight than women with a normal pregnancy. There were no significant differences between the study groups in the gestational age at sample collection (normal pregnancy: median 31.6 weeks (range 20–38.3) vs. FD: median 31 weeks (range 19.0–40.0), P = 0.7).

Table I.  Demographic and clinical characteristics of patients with normal pregnancies and those with fetal demise.

	Normal pregnancy (n = 71)	Fetal demise (n = 65)	P value
Maternal age (years)	24 (21–27)	26 (20–30.5)	0.2
Gravidity*
1	14 (20.3%)	19 (30.2%)	0.2
2–5	45 (65.2%)	31 (59.2%)
≥6	10 (14.5%)	13 (20.6%)
Parity^†
1	39 (55.7%)	43 (67.2%)	0.1
2–5	30 (42.9%)	18 (28.1%)
≥6	1 (1.4%)	3 (4.7%)
Ethnic origin^§
African-Americans	53 (74.6%)	55 (84.6%)	0.6
Caucasian	11 (15.5%)	5 (7.7%)
Hispanic	2 (2.8%)	2 (3.1%)
Asian	2 (2.8%)	1 (1.5%)
Gestational age at blood collection (weeks)	31.6 (27.4–35)	31.0 (24.0–35.2)	0.7
Gestational age at delivery (weeks)	39.6 (38.4–40.6)	31.0 (25.2–35.4)	<0.001
Neonatal birthweight (g)	3330 (3050–3700)	1170 (570–2192.5)	<0.001

Data are presented as median (interquertile range) or numbers (%).

*Normal pregnancy (n = 69); Fetal demise (n = 63).

^†Normal pregnancy (n = 70); Fetal demise (n = 64).

^§Normal pregnancy (n = 68); Fetal demise (n = 63).

In the FD group, 23.1% (15/65) of the patients had hypertensive disorders of pregnancy, of which 9.3% (6/65) had gestational hypertension and 13.8% (9/65) had preeclampsia. Because our group has previously reported the association between preeclampsia and elevated TF and TFPI maternal plasma concentrations, we analyzed the changes in the coagulation profile of the maternal plasma in three steps: (1) all patients with a FD were included as one group; (2) in order to study the effect of gestational hypertension, we subdivided the FD group into two subgroups, patients with a FD without hypertension( n = 50) and patients with a FD with hypertension (n = 15), and both groups were compared with the normal pregnancy group; (3) in the third analysis, we studied the effect of preeclampsia, and the FD group was subdivided in two subgroups: patients with a FD without hypertension(n = 50) and patients with a FD and preeclampsia (n = 9). Patients with gestational hypertension were excluded from this analysis and both groups were compared with the normal pregnancy group.

Changes in the coagulation profile of women with normal pregnancies and patients with a fetal demise

In the normal pregnancy group, there was a positive correlation between maternal plasma immunoreactive TF concentrations and activity (r = 0.327, P = 0.006), but not between maternal plasma TFPI concentration and activity (r = 0.22, P = 0.06).

The median maternal plasma immunoreactive TF concentration and activity did not differ significantly between patients with a FD and women with normal pregnancies (P = 0.6 and P = 0.06, respectively; Table II). In contrast, the median maternal plasma TFPI concentration, but not its activity, was lower in patients with a FD than in women with a normal pregnancy (P < 0.001 and P = 0.093, respectively; Table II).

Table II.  Maternal plasma coagulation parameters of patients with a normal pregnancy and those with a fetal demise.

	Normal pregnancy (n = 71)	Fetal demise (n = 65)	P value
Tissue factor concentration (pg/ml)	345.7 (21.7–2662.2)	298.6 (63.6–1987.0)	0.6
Tissue factor activity (pM)	9.9 (0.7–37.6)	11.4 (5.3–59.3)	0.06
Tissue factor pathway inhibitor concentration (ng/ml)	66.7 (37.4–86.5)	48.5 (13.6–123.7)	<0.001
Tissue factor pathway inhibitor activity (unit/ml)	1.1 (0.7–2.7)	1.2 (0.8–2.5)	0.09
Soluble CD40L (pg/ml)	369.5 (63.5–1848.7)	988.3 (118.0–3818.0)	<0.001
Thrombin–antithrombin III complexes (μg/l)	15.7 (5.2–107, 941.0)	29.3 (6.6–61, 256.8)	<0.001
Prothrombin fragments 1 + 2 (pmol/l)	3, 786.0 (513.4–486, 310.0)	5, 323.5 (343.4–494, 564.0)	<0.001

Data are presented as median (minimum, maximum).

Patients with a FD had a significantly higher median maternal plasma sCD40L than those with a normal pregnancy (P < 0.001). In addition, the median maternal plasma concentrations of TAT III complexes and PF 1 + 2 were higher in patients with a FD than in normal pregnant women (P < 0.001 for both comparisons; Table II).

The effect of preeclampsia on the coagulation profile of patients with fetal demise

The median maternal plasma TF concentration and activity differed significantly among patients with a FD with and without preeclampsia, as well as women with normal pregnancies (Kruskal–Wallis, P < 0.001 and P = 0.001, respectively). Patients with a FD and preeclampsia had a higher median maternal plasma TF concentration and activity than that of women with normal pregnancies (P = <0.001 and P = 0.003, respectively), and than that of patients with a FD without hypertension (P < 0.001 and P = 0.02, respectively) (Figures 1(a) and 1(b)).

Figure 1. (a) Maternal plasma tissue factor (TF) concentrations among women with normal pregnancies (median 345.7 pg/ml, range 21.7–2662.2) and patients with a fetal demise with preeclampsia and those without hypertension (with preeclampsia: median 1167.0 pg/ml, range 287.5–1987.0; without hypertension: median 284.2 pg/ml, range 123.6–851.6); (b) Maternal plasma TF activity among women with normal pregnancies (median 9.9 pM, range 0.7–37.6) and patients with a fetal demise with preeclampsia and those without hypertension (with preeclampsia: median 16.2 pM, range 11.6–59.3; without hypertension: median 10.9 pM, range 5.3–57.2).

The median maternal plasma TFPI concentration and activity differed significantly among patients with a FD with and without preeclampsia and women with normal pregnancies (Kruskal–Wallis, P < 0.001 and P = 0.021, respectively). Patients with a FD and preeclampsia had a higher median maternal plasma TFPI activity, but not TFPI concentration, than women with normal pregnancies (P = 0.005, P = 0.1, respectively). Among the FD group, patients with preeclampsia had a higher median maternal plasma TFPI concentration than those without hypertension, but the median TFPI activity did not differ significantly between the groups (P = 0.003 and P = 0.1, respectively) (Figures 2(a) and 2(b)).

Figure 2. (a) Maternal plasma tissue factor pathway inhibitor (TFPI) concentration among women with normal pregnancies (median 66.7 ng/ml, range 37.4–86.5) and patients with a fetal demise with preeclampsia and those without hypertension (with preeclampsia: median 72.4 ng/ml, range 45.3–123.7; without hypertension: median 45.6 ng/ml, range 13.6–115.2); (b) Maternal plasma TFPI activity among women with normal pregnancies (median 1.1 unit/ml, range 0.7–2.7) and patients with a fetal demise with preeclampsia and those without hypertension (with preeclampsia: median 1.4 unit/ml, range 1.0–1.7; without hypertension: median 1.15 unit/ml, range 0.8–2.5).

The median maternal plasma sCD40L differs significantly between patients with a FD with and without preeclampsia and women with normal pregnancies (Kruskal–Wallis, P < 0.001). However, in the comparison between the groups, the median maternal plasma sCD40L concentration did not differ significantly between patients with a FD with preeclampsia and women with a normal pregnancy (P = 0.1) or those with a FD without hypertension (P = 0.1) (Figure 3).

Figure 3. Maternal plasma sCD40L concentration among women with normal pregnancies (median 369.5 pg/ml, range 63.5–1848.7) and patients with a fetal demise with preeclampsia and those without hypertension (with preeclampsia: median 732.0 pg/ml, range 271.3–1714.0; without hypertension: median 1213.3 pg/ml, range 118.0–3818.0).

The median plasma TAT III complexes and PF 1 + 2 concentration differed significantly among these groups (Kruskal–Wallis, P < 0.001, for both comparisons). However, patients with preeclampsia and FD had a significantly higher median plasma TAT III complexes and PF 1 + 2 concentration than women with a normal pregnancy (P = 0.003 for both comparisons), but not those with FD without hypertension (P = 0.1 and P = 0.25, respectively) (Figures 4 and 5).

Figure 4. Maternal plasma thrombin–antithrombin (TAT) III complexes concentration among women with normal pregnancies (median 15.7 μg/l, range 5.2–107, 941.0) and patients with a fetal demise with preeclampsia and those without hypertension (with preeclampsia: median 105.8 μg/l, range 11.2–2002.8; without hypertension: median 28.9 μg/l, range 6.6–61, 256.8).

Figure 5. Maternal plasma prothrombin fragments 1 + 2 concentration among women with normal pregnancies (median 3786.0 pmol/l, range 513.4–486, 310.0) and patients with a fetal demise with preeclampsia and those without hypertension (with preeclampsia: median 8715.4 pmol/l, range 4320.4–18, 954.2; without hypertension: median 5205.4 pmol/l, range 343.4–494, 564.0).

The association between placental lesions and changes in markers of coagulation, platelet activation and thrombin generation in patients with a normal pregnancy and those with a fetal demise

Histologic examination of the placenta was performed in 15.5% (11/71) of women with a normal pregnancy and 80% (52/65) of patients with a FD. Within the latter group, 82% (41/50) of patients with a FD without hypertension and 73.3% (11/15) of those with a FD and hypertension had histologic examination of their placenta. Among patients with a FD, those with hypertension had a higher rate of villous infarcts, syncytial knots, intervillous fibrin deposition, and acute atherosis of basal plate arteries or decidual arterioles than patients with FD without hypertension (Table III).

Table III.  Placental lesions in patients with a fetal demise according to the presence of hypertensive diseases of pregnancy

Placental histologic findings	Fetal demise without hypertension (n = 41)	Fetal demise with hypertension (n = 11)	P value
Findings consistent with amniotic fluid infection
Chorioamnionitis, maternal response	9 (21.9%)	3 (27.3%)	0.7
Funisitis, fetal response	3 (7.3%)	0	1
Findings consistent with maternal under perfusion
Remote villous infarcts	3 (7.3%)	4 (36.4%)	0.029
Recent villous infarcts	1 (2.4%)	1 (9.1%)	0.4
Increased syncytial knots	3 (7.3%)	4 (36.4%)	0.029
Villous agglutination	2 (4.9%)	3 (27.3%)	0.06
Increased intervillous fibrin	1 (2.4%)	3 (27.3%)	0.026
Distal villous hypoplasia	3 (7.3%)	1 (9.1%)	1
Persistent muscularisation of basal plate arteries	2 (4.9%)	1 (9.1%)	0.5
Mural hypertrophy of decidual arterioles	0	0	–
Acute atherosis of basal plate arteries/decidual arterioles	1 (2.4%)	3 (27.3%)	0.026
Findings consistent with fetal vascular thrombotic occlusive disease
Early villous stromal-vascular karyorrhexis	0	0	–
Exclusively small foci	2 (5%)	1 (9.1%)	0.6
Variable size foci	6 (15%)	0
Fetal thrombotic vasculopathy	3 (7.5%)	1 (9.1%)
Thrombi, large fetal vessels	3 (7.5%)	0	1
Chronic villitis with obliterative fetal vasculopathy	0	1 (9.1%)	0.2

Data are presented as numbers (%).

Table IV.  Demographic and clinical characteristics of patients with normal pregnancies and those with a fetal demise that had amniocentesis

	Normal pregnancy (n = 25)	Fetal demise (n = 36)	P value
Maternal age (years)	24 (21–28)	26 (20–29)	0.4
Gravidity
1	3 (12%)	12 (33.3%)	0.2
2–5	16 (64%)	18 (50%)
≥6	6 (24%)	6 (16.7%)
Parity
1	15 (60%)	26 (72.2%)	0.6
2–5	9 (36%)	9 (25%)
≥6	1 (4%)	1 (2.8%)
Ethnic origin
African-Americans	19 (76%)	30 (83.3%)	0.6
Caucasian	4 (16%)	2 (5.6%)
Hispanic	1 (4%)	2 (5.6%)
Asian	1 (4%)	2 (5.6%)
Gestational age at amniocentesis (weeks)	30.3 (26.4–35.4)	32.5 (26.2–36.7)	0.3
Gestational age at delivery (weeks)	39.3 (38.3–40.3)	32.7 (26.3–38.0)	<0.001

Data are presented as median (interquertile range) or numbers (%).

Among patients with a FD and hypertension, women with placental fetal vascular thrombotic occlusive lesions had a significantly higher median maternal plasma sCD40L concentration than those without these placental findings (P = 0.036). In addition, among patients in the FD group without hypertension, those who had placental lesions associated with maternal under perfusion had a significantly higher PF 1 + 2 than those without these lesions (P = 0.025). Of note, among patients with normal pregnancies there were no associations between the median maternal plasma concentration of the different analytes and a specific placental lesion (Supplementary Table I–III).

Supplementary table I. The association between placental lesion and markers for coagulation in women with normal pregnancy.

Analyte Concentration/activity	Histologic chorioamnionitis and funisitis		Maternal under perfusion		Fetal vascular thrombo-occlusive disease
	With n = 2	Without n = 9	With n = 1	Without n = 10	With n = 3	Without n = 8
TF (pg/ml)	260.7 (175.6–345.7)	440.9 (22.8–911.2)	258.9	416.3 (22.8–911.2)	499.2 (345.7–870.2)	385.3 (22.8–911.2)
TF activity (pM)	16.0 (10.1–22.0)	12.3 (5.7–28.1)	5.7	14.2 (8.9–28.1)	22.0 (9.4–28.1)	10.3 (5.7–26.4)
TFPI (ng/ml)	60.2 (59.7–60.6)	66.7 (41.3–76.6)	66.7	63.6 (41.3–76.6)	60.6 (41.3–76.6)	66.7 (48.5–76.5)
TFPI activity (u/ml)	0.9 (0.9–1.1)	1.1 (0.9–1.8)	1.0	1.1 (0.9–1.7)	1.0 (0.9–1.7)	1.1 (0.9–1.8)
sCD40L (pg/ml)	805.2 (352.3–1258.0	325.0 (101.9–1848.7)	312.3	418.8 (101.9–1848.7)	352.3 (325.0–485.3)	687.0 (101.9–1848.7)
TAT III complexes (μg/L)	17.0 (9.4–24.5)	19.6 (7.7–107941.0)	12.8	20.1 (7.7–107941.0)	24.5 (20.7–72628.9)	14.2 (7.7–107941.0)
PF 1+2 (pmol/L)	5946.1 (1662.1–10230.0)	4282.9 (622.4–486310.0)	8308.4	4242.1 (622.4–486310.0)	10230.0 (3551.9–486310.0)	4242.1 (622.4–311448.0)

*P < 0.05; TF- tissue factor; TFPI – tissue factor pathway inhibitor; TAT- thrombin anti-thrombin III; PF- prothrombin fragments

Supplementary table II. The association between placental lesion and markers for coagulation in women with fetal demise without hypertension.

Analyte Concentration/activity	Histologic chorioamnionitis and funisitis		Maternal under perfusion		Fetal vascular thrombo-occlusive disease
	With n = 8	Without n = 34	With n = 7	Without n = 34	With n = 12	Without n = 29
TF (pg/ml)	322.1 (187.6–851.6)	278.8 (167.9–487.6)	265.4 (167.9–345.6)	294.1 (167.9–851.6)	284.2 (176.8–476.5)	298.5 (167.9–851.6)
TF activity (pM)	9.7 (6.6–14.2)	11.2 (5.3–57.3)	19.4 (5.3–57.2)	10.5 (6.4–52.6)	10.9 (5.3–57.2)	10.8 (6.6–52.7)
TFPI (ng/ml)	44.1 (34.6–48.8)	54.6 (13.6–115.2)	59.7 (34.6–74.1)	47.1 (13.6–115.2)	51.7 (13.6–73.2)	45.6 (29.8–115.2)
TFPI activity (u/ml)	1.3 (1.01–1.7)	1.1 (0.8–2.5)	1.0 (0.9–2.5)	1.2 (0.8–2.0)	1.1 (0.9–1.5)	1.2 (0.8–2.5)
sCD40L (pg/ml)	1634.0 (335.3–2598.0)	1282.2 (118.0–3818.0)	1365.0 (412.0–2256.4)	1282.2 (118.0–3818.0)	1704.0 (118.0–3408.0)	1223.2 (225.0–3818.0)
TAT III complexes (μg/L)	24.9 (13.4–77.8)	29.8 (9.6–61256.8)	38.1 (14.9–704.1)	28.9 (9.6–61259.8)	40.8 (9.6–61256.8)	28.9 (13.4–399.4)
PF 1+2 (pmol/L)	4762.8 (1512.7–22129.5)	6882.5 (968.2–494564.0)	13266.1* (4616.8–19650.5)	5205.4 (968.2–494564.0)	9512.7 (1512.7–494564.0)	5273.1 (968.2–22129.5)

*P < 0.05; TF- tissue factor; TFPI – tissue factor pathway inhibitor; TAT- thrombin anti-thrombin III; PF- prothrombin fragments

Supplementary table III. Supplementary The association between placental lesion and markers for coagulation in women with fetal demise with hypertension.

Analyte Concentration/activity	Histologic chorioamnionitis and funisitis		Maternal under perfusion		Fetal vascular thrombo-occlusive disease
	With n = 3	Without n = 7	With n = 7	Without n = 3	With n = 2	Without n = 9
TF (pg/ml)	1187.0 (765.0–1987.0)	862.0 (287.5–1876.0)	1167.0 (325.8–1987.0)	862.0 (287.5–1458.0)	1577.0 (1167.0–1987.0)	765.0 (287.5–1876.0)
TF activity (pM)	17.0 (14.7–22.0)	16.2 (11.7–59.3)	16.5 (12.2–59.1)	12.2 (11.7–22.6)	17.1 (12.2–22.0)	16.2 (8.0–59.3)
TFPI (ng/ml)	98.5 (72.4–119.3)	67.6 (45.3–123.7)	72.4 (47.6–119.3)	67.6 (45.3–123.7)	94.6 (72.4–116.8)	67.6 (37.8–123.7)
TFPI activity (u/ml)	1.6 (1.4–1.7)	1.4 (1.0–2.2)	1.6 (1.2–2.2)	1.2 (1.0–1.4)	1.4 (1.36–1.43	1.4 (1.0–2.2)
sCD40L (pg/ml)	942.1 (732.0–1714.0)	356.4 (217.3–1638.0)	732.0 (217.3–1714.0)	356.4 (268.0–872.0)	1676.0* (1638.0–1714.0)	618.0 (217.3–1268.0)
TAT III complexes (μg/L)	56.0 (11.2–105.8)	303.6 (19.4–2002.8)	560 (11.2–2002.8)	303.6 (21.7–374.9)	15.3 (11.2–19.4)	105.8 (13.4–2002.8)
PF 1+2 (pmol/L)	8715.4 (4347.7–18954.2)	8999.6 (4320.4–17983.7)	13936.2 (4347.7–8954.2)	7753.6 (4320.4–8999.6)	4706.5 (4347.7–5060.2)	8999.6 (758.6–18954.2)

*P < 0.05; TF- tissue factor; TFPI – tissue factor pathway inhibitor; TAT- thrombin anti-thrombin III; PF- prothrombin fragments

Changes in the median amniotic fluid TF concentrations and activity as well as the concentrations of thrombin–antithrombin III complexes

The demographic and clinical characteristic of the patients are presented in Table IV.

The TFPI concentrations, TFPI activity and PF 1 + 2 assays had severe matrix effect in AF and were excluded from the study. Of note, we could not detect immunoreactive sCD40L in the AF.

Patients with a FD had higher median AF–TF concentration and activity than those with normal pregnancies (P < 0.001, for both comparisons; Figures 6(a) and 6(b), respectively). Moreover, among patients with a FD there was a significant correlation between the AF–TF concentrations and activity (r = 0.88, P < 0.0001). The median AF–TAT III complexes concentration did not differ significantly between the groups (normal pregnancy: median: 66.3 μg/l, range 11.4–2265.4 vs. FD: median: 59.3 μg/l, range: 13.6–15,425.3; P = 0.7).

Figure 6. (a) Amniotic fluid tissue factor concentration among women with normal pregnancies (median 3710.4 pg/ml, range 2198.8–6268) and patients with a fetal demise (median 8535.4 pg/ml, range 2208.2–125,990.0); (b) Amniotic fluid tissue factor activity among women with normal pregnancies (median 28.4 pM, range 10.2–84.9) and patients with a fetal demise (median 81.6 pM, range 7.2–1603.4).

The three patients in the FD group with placental abruption had a higher median AF–TAT III complexes concentration than women with normal pregnancies (FD with abruption: median 754.1 μg/l range 398.8–10,573.8; P = 0.015) and patients with FD without abruption (FD without abruption: median 55.5 μg/l range 13.6–15,425.3; P = 0.012).

Discussion

Principal findings of the study

(1) The median maternal plasma sCD40L (an indicator of in vivo platelet activation) concentration was higher in patients with FD than that of normal pregnant women; (2) similarly, patients with a FD had a higher median maternal plasma concentration of TAT III complexes and PF 1 and 2, indices of increase thrombin generation in vivo, than those of women with a normal pregnancy; (3) a higher median AF–TF concentration and activity was detected in patients with a FD than in women with normal pregnancies. However, this was not detected in maternal plasma. Collectively, these results suggest that mothers with a fetal death have a state of platelet activation, and increased thrombin generation, even in the absence of DIC.

Evidence of maternal platelet activation in fetal death

Soluble CD40L, a marker for platelet activation [43-45], participates in platelet aggregation and platelet–leukocyte conjugation [46]. Upon activation, platelets express CD40L on their membrane (from which it is subsequently cleaved by CD40), and the soluble form of CD40L (sCD40L) can be measured in the plasma [47],[48]. Activated platelets are the source of more than 90% of soluble sCD40L in the plasma [47], and this cytokine has been proposed as a measurable marker of platelet activation [43],[44].

The finding of a higher median plasma concentration of sCD40L in patients with a FD is novel and can be interpreted as indicating in vivo platelet activation in these patients. There is a solid body of evidence supporting platelet activation in preeclampsia and fetal growth restriction [49-73]. Indeed, elevated maternal plasma and platelet expression of markers for platelet activation, including P-selectin (CD62p) [52-54],[61],[63],[65],[74],[75], Annexin V [53],[57],[76], platelet factor 4 [29],[49],[77-81], and sCD40L concentrations [82-85] were reported in these conditions. Moreover, we have recently reported that the median maternal plasma sCD40L concentration is elevated in patients with preterm labor and intact membranes[86]; of note, all these obstetrical syndromes were associated with elevated median maternal plasma TAT III complexes concentration [21],[30]. There is, however, a paucity of data about the degree and mechanisms responsible for platelet activation in cases of fetal death.

Normotensive patients with a FD had a higher median plasma sCD40L concentration than women with a normal pregnancy as well as than that of patients with preeclampsia and a FD. Thus, patients with a FD have platelet activation that cannot be accounted by hypertensive disorders of pregnancy. Moreover, this finding suggests that in patients with a dead fetus, activated platelets may play a role in the generation of thrombin as reflected by the elevated median plasma concentration of PF 1 + 2 and TAT complexes. However, the association between platelet activation and higher indices of thrombin generation is complex, and each system can be activated by itself, or by the other [87-95].

Several possible mechanisms may contribute to platelet activation in normotensive patients with a FD: (1) platelet activation can be mediated by local/intervillous activation of the coagulation cascade, which may be initiated by the loss of inhibitory mechanisms of TF activation in the trophoblasts. This can lead to an increased thrombin generation, which may activate platelets through the protease-activated receptors (PAR) [96-103], and the glycoprotein Ib receptors [102-105] that also amplify the effect of thrombin on PAR-1 [106]. The activation of maternal platelets leads to an increased median plasma sCD40L concentration and to the release of partially-activated factor V from α granules onto the platelet surface [107], as well as the activation of the intrinsic pathway of coagulation through factors IX and XI [108-111]. (2) Platelet activation can occur as a consequence of maternal inflammation associated with active villous destruction by infiltrating neutrophils. This has been reported in RU486 treated cynomolgus monkeys with a FD [112]. Furthermore, in women with a dead fetus, neutrophils may be activated by the elevated maternal plasma C5a concentrations [113] and subsequently, these neutrophils can activate platelets by secreting cathepsin G [114]. Collectively, the increased platelet activation reported here in patients with a FD may further propagate the increase in indices for thrombin generation associated with this obstetrical syndrome.

Changes in indices for thrombin generation, thrombin–antithrombin complexes and prothrombin fragments 1 + 2, in patients with a fetal demise

The finding of higher median maternal plasma concentrations of TAT III complexes and PF 1 + 2 in patients with a FD without maternal DIC is novel, as they represent indices of increased in vivo thrombin generation. This observation is in keeping with previous reports of an increased thrombin generation in several obstetrical syndromes, including pre-eclampsia [26-32],[115], a small-for-gestational age neonate[30],[34],[35], preterm labor[21],[36], and preterm PROM [21],[37]. Although DIC and hypofibrinogenemia in patients presenting with a retained dead fetus were reported already in the early 1950's [1-6], the maternal plasma thrombin generation in women with a dead fetus without DIC was not studied. This is the first report of elevated median maternal plasma concentrations of TAT III complexes and PF 1 + 2, indices for in vivo thrombin generation, in these patients.

The underlying mechanisms leading to excessive thrombin generation in patients with a fetal death have not been elucidated. Abnormalities of the placenta, specifically vascular placental lesions, may be associated with the above observation. Indeed, among patients with a FD without hypertension, those who had vascular placental lesions consistent with maternal under perfusion had a higher median plasma PF 1 + 2 concentration than those without these lesions. Moreover, the human trophoblast has properties of endothelial cells and can regulate the degree of activation of the coagulation cascade in the intervillous space [116],[117]. Unlike the endothelium of other organs, the trophoblast constantly presents the active placental isoform of TF on its surface [15],[116-120]. This isoform is larger than recombinant TF and has a higher affinity for FVIIa [121],[122], which may lead to increased activation of the coagulation cascade. Consequently, in dead fetuses, the trophoblast may no longer suppress the procoagulant activity of the placental isoform of TF, leading to subclinical activation of the hemostatic system and increased thrombin generation in the maternal plasma, even in the absence of DIC.

Changes in amniotic fluid and maternal plasma tissue factor concentration and activity in patients with a fetal demise

This is the first study to report that patients with a FD have an increased median AF–TF concentration and activity. This is relevant because the release of TF has traditionally been considered responsible for the initiation of the maternal consumptive coagulopathy in cases of FD [1]. High TF concentrations have been previously detected in the fetal membranes (mainly the amnion) and AF of normal pregnant women [16-19]. Moreover, the AF–TF concentrations were higher than that of the maternal plasma [19]. In our study, the median AF–TF concentration in normal pregnant women was 10 fold higher than in maternal plasma. Furthermore, the median AF–TF concentration and activity in women with a FD without DIC, was higher than those in the maternal plasma of normal pregnant women (30 fold and 7–8 fold, respectively).

In contrast to the changes observed in AF, the median maternal plasma TF concentration and activity did not differ between patients with a FD without hypertension and those with a normal pregnancy. This raises a question regarding the role of circulating immunoreactive TF in the pathophysiology of fetal death syndrome.

It is unclear whether the maternal plasma TF concentration reflects activation of the coagulation cascade or systemic maternal inflammation. Indeed, the procoagulant activity of plasma immunoreactive TF (blood-borne TF) of non-pregnant patients [123-128] is controversial. Blood-borne TF has very little or no procoagulant activity, and only the administration of exogenous active TF generates a whole blood and plasma clot after the inhibition of the contact factor (factor XIIa) [128]. Others [124-127] have proposed that blood-borne TF does not initiate the coagulation cascade but rather, propagates clot formation by attaching to activated platelets, further enhancing the coagulation process. However, our finding of a positive correlation between maternal plasma TF concentration and its activity in women with a normal pregnancy suggests that circulating TF has procoagulant properties that may be part of the hypercoagulable state of pregnancy.

The absence of a significant difference in the median maternal plasma immunoreactivity and activity of TF between women with normal pregnancies and normotensive patients with a FD raises a question regarding the validity of the proposed role of tissue thromboplastin in the mechanism leading to DIC in women with fetal death syndrome. When these mechanisms were proposed, there was no direct assay to measure the presence and concentrations of placental TF in the maternal circulation of patients with a FD. However, this view was supported by the experiment reported by Schneider [129] which demonstrated that the intravenous injection of placental extracts into mice leads to the death of the animal through DIC, which can be prevented by the administration of heparin [129]. The author identified thromboplastin as the causative agent through its effect on the clotting time and its chemical properties, and measured its activity by the one-stage prothrombin-time method [129]. Currently, early detection of fetal death by ultrasound and the practice of induction of labor have almost abolished the development of fetal death syndrome in western societies. Thus, the verification of this hypothesis by advanced immunoreactive and functional assays is less feasible. Indeed, we did not have any patient with maternal DIC in our study group. Therefore, we suggest that among patients presenting with a FD without preeclampsia and without DIC, blood-borne TF may have little or no role in the increased maternal plasma indices for thrombin generation detected in these patients, and the role of the increased AF–TF concentration and activity awaits further study.

Preeclampsia and tissue factor concentration and activity

In the FD group, preeclampsia was associated with a higher median maternal plasma TF concentration and activity than that of patients with a normal pregnancy. Moreover, the median TF concentration of patients with preeclampsia was also higher than in patients with a FD without hypertension. These findings are consistent with previous studies [130],[131], suggesting that elevated TF immunoreactivity and activity may be associated with the pathophysiologic process leading to preeclampsia, rather than a consequence of the fetal death.

The hemostatic system in patients having a fetal demise without DIC

FD is syndromic in nature and results from different underlying mechanisms, which can be acute (i.e. placental abruption [132-140], penetrating trauma [141], prolapse of umbilical cord [142],[143]) or chronic (placental infarcts [144-160], infection [145],[146],[161-163], chromosomal or anatomic malformation [145],[146],[161],[163],[164]). The profile of maternal plasma and AF coagulation presented herein may be associated with some of the underlying mechanisms leading to fetal death.

In patients with preeclampsia and a FD, the median TF plasma concentration and activity were higher than those of normal pregnant women, suggesting activation of the coagulation cascade that is probably due to preeclampsia rather than the dead fetus. On the other hand, patients with a FD without hypertension had an elevated median plasma sCD40L concentration, which indicates increased platelet activation.

All the patients included in our study did not have DIC, suggesting that the increased platelet activation and elevated indices of thrombin generation reflect subclinical activation of the hemostatic system. Evidence in support of this view is derived from the study by Brummel et al. [165], who calculated that at least 26 ± 6 nM of TAT III complexes are needed in order to reach the propagation phase of thrombin generation (909 μg/l of TAT III complexes). In the current study, translation of the median TAT III complexes concentration into molar quantity revealed that women with normal pregnancies had median maternal plasma of TAT complexes 0.17 nM, which increased to 0.31 nM in the case of patients with a FD, and to 0.55 nM among those with a FD and hypertension. Moreover, the highest concentration was observed in the AF of patients with placental abruption, reaching 7.96 nM of TAT III complexes.

Consequently, these findings suggest that although women with a FD have higher thrombin generation than women with normal pregnancies, they may still be at the initiation phase of thrombin generation. Thus, in the absence of an acute cause of fetal death, like abruption, that will lift the dam and unleash the waterfall of the coagulation cascade, leading to uncontrolled thrombin generation, maternal consumption coagulopathy and DIC; the activation of the hemostatic system is subclinical and well tolerated by patients with a dead fetus at least for a limited time period.

Summary

We report the maternal plasma and AF elements of the initiation and propagation steps of the coagulation cascade, as well as indices of thrombin generation, which is the outcome of the activity of the formers and the inhibitor of initiation step. The question regarding the role of TF in the development of maternal DIC remains unanswered and can not be addressed by our study, because none of the patients had DIC and the current medical practice of induction of labor near to the diagnosis of a dead fetus seems to prevent this maternal complication. On the other hand, we report here for the first time that patients with a FD have increased platelet activation and indices of thrombin generation even before the development of DIC, suggesting subclinical activation of the hemostatic system.

Acknowledgement

This research was supported (in part) by the Perinatology Research Branch, Division of Intramural Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH, DHHS.