A link between a hemostatic disorder and preterm PROM: a role for tissue factor and tissue factor pathway inhibitor

Abstract

Objective. Vaginal bleeding is a risk factor for preterm PROM (PPROM). A disorder of decidual hemostasis has been implicated in the genesis of PROM. Indeed, excessive thrombin generation has been demonstrated in PPROM both before and at the time of diagnosis. Decidua is a potent source of tissue factor (TF), the most powerful natural pro-coagulant. A decidual hemostatic disorder may link vaginal bleeding, PPROM and placental abruption. This study was conducted to determine the behaviour of maternal TF and its natural inhibitor, the tissue factor pathway inhibitor (TFPI) in PPROM.

Methods. This cross-sectional study included women with PPROM (n = 123) and women with normal pregnancies (n = 86). Plasma concentrations of TF and TFPI were measured by a sensitive immunoassay. Non-parametric statistics were used for analysis.

Results. (1) The median maternal plasma TF concentration was significantly higher in patients with PPROM than in women with normal pregnancies (median: 369.5 pg/mL; range: 3.27–2551 pg/mL vs. median: 291.5 pg/mL; range: 6.3–2662.2 pg/mL respectively, p = 0.001); (2) the median maternal TFPI plasma concentration was significantly lower in patients with PPROM than in women with normal pregnancies (median: 58.7 ng/mL; range: 26.3–116 ng/mL vs. median: 66.1 ng/mL; range: 14.3–86.5 ng/mL respectively, p = 0.019); (3) there was no correlation between the plasma concentration of TF and TFPI and the gestational age at sample collection; and (4) among patients with PPROM there was no association between the presence of intra-amniotic infection or inflammation and median plasma concentrations of TF and TFPI.

Conclusions. (1) Patients with PPROM have a higher median plasma concentration of TF and a lower median plasma concentration of TFPI than women with normal pregnancies. (2) These findings suggest that PPROM is associated with specific changes in the hemostatic/coagulation system.

Keywords:

• Coagulation
• inflammation
• pregnancy
• preterm delivery
• thrombin
• thrombosis
• prematurity
• prelabor rupture of membranes

Introduction

Preterm prelabour rupture of membranes (PROM) is associated with placental vascular lesions including a failure of physiologic transformation of the spiral arteries [1], atherosis and decidual vasculopathy (i.e. fibrinoid necrosis and thrombosis of decidual vessels) [1,2]. Moreover, women with preterm PROM have a higher median maternal plasma concentration of thrombin–antithrombin (TAT) complexes than women with normal pregnancies [3,4]. Thus, the activation of the coagulation cascade that results in an increased thrombin generation may be one of the mechanisms of disease leading to preterm PROM.

The activation of the coagulation system in the placental and maternal compartment of patients with preterm PROM can result from the following underlying mechanisms of disease: (1) decidual hemorrhage that leads to a retro-placental clot formation [5]; (2) intra-amniotic infection which can induce decidual bleeding and sub-clinical abruption [6], as well as increased intra-amniotic TAT complexes [7]; and (3) an increased maternal systemic inflammatory response [8] that may activate the extrinsic pathway of coagulation due to the expression and release of tissue factor (TF) by activated monocytes [9]. These mechanisms result in an increased thrombin generation, which has been associated with the following: (1) stimulation of decidual cell secretion of matrix metalloproteinases (MMP) (i.e. MMP-1 [10] and MMP-3 [11]) that can degrade the extracellular matrix of the chorioamniotic membranes; and (2) myometrial activation and uterine contractions generation that may lead to preterm labour with or without rupture of membranes and a subsequent preterm delivery [12-14].

Although thrombin is generated as a consequence of activation of the coagulation cascade, TF, the most powerful natural pro-coagulant, is abundant in the uterine decidua in the normal state [15,16]. Tissue factor is part of an efficient hemostatic mechanism in the uterine wall, which is activated in the course of normal pregnancy during implantation [17] and after delivery [18]. However, this hemostatic mechanism may also be activated due to pathological decidual bleeding in pregnancies complicated by placental abruption [5],[19] and intra-amniotic infection [6].

The main physiological inhibitor of the TF pathway is tissue factor pathway inhibitor (TFPI), which is a glycoprotein comprised of three Kunitz domain [20] that are specific inhibitors of trypsin-like proteinases [21]. The mean maternal plasma concentration of total TFPI (TFPI-1 and TFPI-2) have been reported to increase during pregnancy until 20 weeks of gestation, to remain relatively constant until term [22] and to decrease during labour [23]. There are two types of TFPI: (1) TFPI-1 is the more prevalent form in the non-pregnant state in the maternal circulation and can also be found in the fetal blood, platelets, endothelial cells and other organs [24,25]; and (2) TFPI-2, the major form of TFPI in the placenta [26-31], is also known as Placental Protein 5 (PP5) [32,33]. During pregnancy, the maternal plasma concentration of TFPI-2 increases gradually, reaches a plateau at 36 weeks and subsides after delivery [33-38].

Pregnancy complications are associated with changes in the maternal plasma and placental concentration of TF and TFPI. Maternal plasma concentrations of TF and free TFPI are higher in women with pre-eclampsia than in patients with a normal pregnancy [39-41]. Placental TF concentrations and mRNA expression are higher in patients with severe pre-eclampsia than in those with a normal pregnancy [42]. In contrast, placentas of patients with gestational vascular complications (pre-eclampsia, eclampsia, fetal growth restriction, placental abruption and fetal demise) have lower placental concentrations of total TFPI and mRNA expression than in those of patients with a normal pregnancy [43].

The increased thrombin generation in the maternal plasma of patients with preterm PROM [3,4] reflects the activation of the coagulation cascade; however, it is not clear whether this activation is associated with changes in TF and TFPI plasma concentrations as well. Thus, the objective of this study is to determine the changes in the maternal plasma concentrations of TF and its natural inhibitor TFPI in women with preterm PROM.

Methods

Study design

This cross-sectional study included patients in the following groups: (1) patients with preterm PROM (n = 123), and (2) women with normal pregnancies (n = 86). Patients with multiple pregnancies or fetuses with congenital and/or chromosomal anomalies were excluded.

Samples and data were retrieved from the bank of biological samples and clinical databases. Many of these samples have been previously used 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 women provided a written informed consent prior to the collection of maternal blood. 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 utilisation of the samples for research purposes.

Definitions

Women were considered to have a normal pregnancy if they did not have obstetrical, medical, or surgical complication of pregnancy and delivered a term neonate of appropriate birth weight for gestational age without complications. Preterm PROM was diagnosed by a sterile speculum examination demonstrating the pooling of amniotic fluid in the vagina (with nitrazine and ferning tests when necessary) occurring <37 weeks of gestation in the absence of labour.

Amniotic fluid collection was performed by trans-abdominal amniocentesis under ultrasonographic guidance in order to determine the microbiologic state of the amniotic cavity in a subset of patients at the discretion of the treating physician. Amniotic fluid was transported to the laboratory in a capped plastic syringe and cultured for aerobic, anaerobic bacteria, as well as for genital Mycoplasmas. White blood cell (WBC) count, glucose concentration and Gram stain for microorganisms were performed in amniotic fluid shortly after collection. Intra-amniotic infection was defined by the presence of amniotic fluid cultures that were positive for microorganisms and inflammation by an amniotic fluid WBC count ≥100 cells. The results of the amniotic fluid analyses were used for clinical management. Small for gestational age (SGA) neonate was defined as birthweight below the 10th percentile [44].

Placental histopathologic examinations

Placental tissue samples were taken by systematic random sampling and subsequently fixed in 10% neutral buffered formalin overnight and embedded in paraffin. Five millimetre-thick paraffin sections were stained with hematoxylin and eosin, and examined using bright-field light microscopy. Histopathologic examinations were performed by three pathologists who were blinded to the clinical information. The placental histologic lesions were classified according to a diagnostic schema proposed by Redline et al. [45].

Blood samples collection

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

Human TF immunoassays

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

Human TFPI 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 a murine anti-TFPI monoclonal as the capture antibody. This capturing antibody is directed against the Kunitz-1 domain of the TFPI molecule; therefore, it detects both TFPI-1 and TFPI-2 and measures 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 is ∼10 ng/mL. The correlation between TFPI antigen concentrations and functional activity is ∼r² = 0.785.

Statistical analysis

Plasma concentrations of TF and TFPI were not normally distributed. Thus, a Mann–Whitney U test was used for comparisons of continuous variables and a Chi-square test was used to compare categorical variables. The Spearman's rho test was used to detect a correlation between the maternal plasma TF and TFPI concentrations and the gestational age at blood collection in women with normal pregnancies. A p-value <0.05 was considered statistically significant. The statistical package employed was SPSS version 12 (SPSS, Chicago, IL).

Results

Demographic characteristics and changes in maternal plasma TF and TFPI concentrations during normal pregnancy

Table I displays the demographic and clinical characteristics of the study groups. Patients with preterm PROM had a lower median gestational age at blood sample collection and at delivery, as well as a lower median neonatal birth weight in comparison with the control group. There was no significant correlation between the maternal plasma TF and TFPI concentrations and the gestational age at blood sample collection in patients with normal pregnancy (TF: r = −0.009, p = 0.94; TFPI: r = 0.157, p = 0.15) (Figure 1).

Figure 1. The correlations between maternal plasma tissue factor (TF) (1a) and tissue factor pathway inhibitor (TFPI) (1b) concentrations and gestational age at blood collection.

Table I.  Demographic and clinical characteristics of the study population

	Normal pregnancy (n = 86)	Preterm PROM (n = 123)	p Value
Maternal age (years)	24 (21–27)	24 (20.25–31)	0.09
Gravidity*
1	18 (21.4%)	18 (14.6%)	0.3
2–5	53 (63.1%)	78 (63.4%)
≥6	13 (15.5%)	27 (22%)
Parity^†
1	46 (54.1%)	59 (48.8%)	0.1
2–5	38 (44.7%)	53 (43.8%)
≥6	1 (1.2%)	9 (7.4%)
Ethnic origin^‡
African-American	67 (80.7%)	105 (85.4%)	0.1
Caucasian	11 (13.3%)	16 (13.0%)
Hispanic	2 (2.4%)	1 (0.8%)
Asian	3 (3.6%)	1 (0.8%)
Gestational age at blood collection (weeks)	31.1 (27.4–35)	30.1 (26–32.5)	0.024
Gestational age at delivery (weeks)	39.6 (38.4–40.7)	31.6 (27.2–33.1)	<0.0001
Induction of labour	16 (22.9%)	49 (42.6%)	0.006
Cesarean delivery^§	24 (33.8%)	34 (27.9%)	0.5
Neonatal birthweight	3342.5 (3050–3700)	1580 (957–2040)	<0.0001

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

PROM, prelabour rupture of membranes.

*Normal pregnancy (n = 84).

^†Normal pregnancy (n = 85).

^‡Normal pregnancy (n = 68).

^§Normal pregnancy (n = 71); Preterm PROM (n = 122).

Changes in maternal plasma TF and TFPI concentrations among patients with preterm PROM and their association with intra-amniotic infection/inflammation

Of the 86 patients in the normal pregnancy group, 91.9% (79/86) had detectable immunoreactive TF in the plasma. Maternal plasma TF concentrations were significantly higher in women with preterm PROM than those with a normal pregnancy (median: 369.5 pg/mL; range: 3.3–2551 pg/mL vs. median: 291.5 pg/mL; range: 6.3–2662.2 pg/mL respectively, p = 0.001) (Figure 2). In contrast, maternal plasma TFPI concentrations were significantly lower in women with preterm PROM than those with a normal pregnancy (median: 58.7 ng/mL; range: 26.3–116 ng/mL vs. median: 66.1 ng/mL; range: 14.3–86.5 ng/mL respectively, p = 0.019) (Figure 3). Patients with preterm PROM had a significantly lower median maternal plasma TFPI/TF ratio than those with normal pregnancies (median: 142.8; range: 21.8–16422 vs. median: 221.5; range: 25.4–3355.3l respectively, p < 0.0001).

Figure 2. Comparison of maternal plasma tissue factor (TF) concentrations between women with normal pregnancy (n = 79) and patients with preterm PROM (n = 123). Tissue factor plasma concentrations were significantly higher in patients with preterm PROM than in women with a normal pregnancy (median: 369.5 pg/mL; range: 3.27–2551 pg/mL vs. median: 291.5 pg/mL; range: 6.3–2662.2 pg/mL, respectively; p = 0.001).

Figure 3. Comparison of maternal plasma tissue factor pathway inhibitor (TFPI) concentrations between women with normal pregnancy (n = 86) and patients with preterm PROM (n = 123). TFPI plasma concentrations were significantly lower in patients with preterm PROM than in women with a normal pregnancy (median: 58.7 ng/mL; range: 26.3–116 ng/mL vs. median: 66.1 ng/mL; range: 14.3–86.5 ng/mL, respectively; p = 0.019).

Sixty-eight patients in the preterm PROM group underwent amniocentesis, of which 54.4% (37/68) had intra-amniotic infection/inflammation. Isolated intra-amniotic inflammation was diagnosed in 16.2% (11/68) of the cases. Positive amniotic fluid cultures for microorganisms were detected in 38.2% (26/68) of patients with preterm PROM (the frequency of the specific microorganisms is presented in Table II).

Table II.  Microorganisms isolated from the amniotic cavity in patients with preterm PROM who underwent trans-abdominal amniocentesis

Microorganism	n (%)
Ureaplasma urealyticum	6 (23.07%)
Mycoplasma hominis	2 (7.6%)
Niesseria gonorrhea	2 (7.6%)
Streptococcus agalactiae	2 (7.6%)
Mixed flora	10 (38.5%)
Others	4 (15.4%)

Table III displays the changes in the median maternal plasma TF and TFPI concentrations according to the presence or absence of intra-amniotic infection and/or inflammation. The patients in both subsets of the preterm PROM group (with and without intra-amniotic infection/inflammation) had a significantly higher median maternal plasma TF concentration than in women with normal pregnancies. There were no significant differences in the median maternal plasma TF and TFPI concentrations between preterm PROM patients with and without intra-amniotic infection and/or inflammation (Table IV).

Table III.  Maternal plasma concentrations of tissue factor and tissue factor pathway inhibitor in patients with preterm PROM with and without intra-amniotic infection and/or inflammation in comparison to women with normal pregnancy

	Preterm PROM without IAI (n = 31)	Preterm PROM with IAI (n = 37)	p Value	Normal pregnancy (n = 86)	p Value*
Tissue factor (pg/mL)^†	423.5^‡ (3.3–899.4)	369.4^‡ (105.3–2551.0)	0.9	291.5 (6.3–2662.2)	0.026
Tissue factor pathway inhibitor (ng/mL)	58.7 (31.2–116.0)	59.7 (37.8–83.6)	0.3	66.1 (14.3–86.5)	0.23

Data presented as median (minimum, maximum).

IAI, intra-amniotic infection/inflammation; PROM, prelabour rupture of membranes.

*Kruskal Wallis test.

^†Normal pregnancy (n = 79).

^‡p < 0.05 in comparison to normal pregnancy group.

Table IV.  Placental lesions in patients with preterm PROM

Placental histologic findings	Preterm PROM (n = 101)
Finding consistent with amniotic fluid infection
Chorioamnionitis, maternal response	59 (58.4%)
Chorioamnionitis, fetal response	49 (48.5%)
Findings consistent with maternal underperfusion
Remote villous infarcts	1 (0.9%)
Recent villous infarcts	–
Increased syncytial knots	5 (4.95)
Villous agglutination	–
Increased intervillous fibrin	1 (0.9%)
Decreased placental weight	–
Distal villous hypoplasia	3 (2.97%)
Persistent muscularization of basal plate arteries	6 (5.9%)
Mural hypertrophy of decidual arterioles	1 (0.9%)
Acute atherosis of basal plate arteries/decidual arterioles	–
Findings consistent with fetal vascular thrombo-occlusive disease
Early villous stromal-vascular karyorrhexis	–
Exclusively small foci	1 (0.9%)
Variable size foci	–
Fetal thrombotic vasculopathy	1 (0.9%)
Thrombi, large fetal vessels	–
Intimal fibrin cushions, large fetal vessels	–
Fibromuscullar sclerosis, Intermediate size fetal vessels	–
Chronic villitis with obliterative fetal vasculopathy	4 (3.96%)

PROM, prelabor rupture of membranes.

Changes in maternal plasma TF and TFPI concentrations among patients with preterm PROM according to their placental lesions and the occurrence of vaginal bleeding during pregnancy

Placental histology was available in 82.11% (101/123) of patients with preterm PROM. Of those, 58.4% (59/101) had histologic chorioamnionitis of maternal response and 48.5% (49/101) had histologic chorioamnionitis of fetal response (i.e. umbilical phlebitis, necrotising funisitis). Persistent muscularisation of the basal plate arteries was the most frequent vascular placental lesion (Table IV). The placental histologic findings were not associated with significant changes in the median maternal plasma concentrations of TF and TFPI.

Vaginal bleeding during pregnancy was documented in 21 preterm PROM patients. The median maternal plasma TFPI concentration was significantly lower in preterm PROM patients without vaginal bleeding than in women with normal pregnancies; conversely, no difference was observed in the median maternal plasma TFPI concentration between patients with preterm PROM and vaginal bleeding and those with normal pregnancies (Table V). In addition, no significant differences were found in the median maternal plasma TF and TFPI concentrations in preterm PROM patients with and without vaginal bleeding during pregnancy [TF: median 412.6 pg/mL, range (256–2551) vs. median 369.5 pg/mL, range (3.3–1879.4), p = 0.5, respectively; and TFPI: median 56.8 ng/mL, range (31.4–74.2) vs. median 58.7 pg/mL, range (26.3–116), p = 0.6, respectively] (Table V).

Table V.  Maternal plasma concentrations of tissue factor and tissue factor pathway inhibitor in patients with preterm PROM with and without vaginal bleeding during pregnancy in comparison to women with normal pregnancy

	Preterm PROM without vaginal bleeding (n = 100)	Preterm PROM with vaginal bleeding (n = 21)	p Value	Normal pregnancy (n = 86)	p Value*
Tissue factor (pg/mL)^†	369.5^‡ (3.3–1879.4)	412.6^‡ (256–2551)	0.5	291.5 (6.3–2662.2)	0.003
Tissue factor pathway inhibitor (ng/mL)	58.7^‡ (26.3–116)	56.8 (31.4–74.2)	0.6	66.1 (14.3–86.5)	0.049

Data presented as median (range).

IAI, intra-amniotic infection/inflammation; PROM, prelabour rupture of membranes.

*Kruskal Wallis test.

^†Normal pregnancy (n = 79).

^‡p < 0.05 in comparison to normal pregnancy.

Comments

Principal findings of the study

(1) Women with preterm PROM have significantly higher median plasma TF concentrations than those with normal pregnancies; (2) the median plasma TFPI concentration was significantly lower in patients with preterm PROM than in those with normal pregnancies; (3) there were no significant differences in maternal plasma TF or TFPI concentrations in patients with preterm PROM with or without intra-amniotic infection/inflammation; and (4) no significant correlation was observed between maternal plasma TF or TFPI concentrations and gestational age at time of blood collection.

What are the changes in TF during pregnancy?

TF is essential for the maintenance of pregnancy and about 90% of TF null mice embryo die at embryonic day 10.5 [46-49] due to abnormalities in yolk sack vasculature [46]. In humans, a substantial increase in decidual and myometrial TF expression and concentrations was reported during the luteal phase of the menstrual cycle [15,16],[50,51], as well as during early and late stages of pregnancy [15,16],[50-52]. Similarly, high TF concentrations have been detected in the fetal membranes (mainly the amnion) and amniotic fluid [23],[53-55]. In contrast, the maternal plasma TF concentrations do not change significantly during normal pregnancy in comparison to non-pregnant women [56,57], though women in labour at term have significantly higher maternal plasma TF concentrations than non-pregnant women [23]. Of interest, TF expression on monocytes is decreased throughout pregnancy, but returns to the normal state by the third day post-partum [58,59].

Why does TF increase in patients with preterm PROM?

Our observation that women with preterm PROM have a significantly higher median maternal plasma concentration of TF is novel. Possible sources for the elevated TF in the maternal plasma could be from decidual activation that is prominent among patients with preterm PROM and from monocyte activation that is associated with the moderate systemic maternal inflammatory state reported in these patients [8]. Both may lead to increased activation of coagulation cascade and higher thrombin generation that can propagate the rupture of membrane through MMP activation [10,11],[60].

TF and activation of coagulation cascade

The higher median maternal plasma concentration of TAT complexes [3,4] along with the higher rate of placental aggregated vascular lesions (atherosis, fibrinoid necrosis of decidual vessels, decidual vessels thrombosis and fetal thrombotic vasculopathy) [1,2],[61] observed in patients with preterm PROM support the suggested increased activation of the coagulation cascade and thrombin generation in these patients.

It has been proposed that a high expression of TF by the decidua and myometrium during normal pregnancy is needed for tight hemostatic control that may prevent decidual bleeding during blastocyst implantation and trophoblast invasion at the early stages of pregnancy, as well as to sustain an adequate uterine hemostasis during the post-partum period [17,18]. Indeed, an increased decidual expression of TF (mRNA and protein) is reported already during the luteal phase of the menstrual cycle and later during pregnancy [62]. Moreover, preterm PROM is associated with an increased activation of the decidual component of the common pathway of parturition [63]. Thus, in pregnancies complicated by abnormal placentation or intrauterine infection, decidual bleeding may lead to a higher expression of TF and activation of the coagulation cascade, resulting in increased thrombin generation. The latter has uterotonic properties that may generate uterine contractions that could initiate labour [12-14]. Moreover, thrombin can mediate the activation of MMP-1 [11], MMP-3 [10] and MMP-9 [60] that can digest components of the extracellular matrix, weaken the choriamniotic membranes and predispose to preterm PROM.

TF and increased maternal intravascular inflammatory response

The mechanisms described above are localised to the maternal-fetal interface. The lack of association between median maternal plasma TF concentrations and the presence of intra-amniotic infection/inflammation or vaginal bleeding in patients with preterm PROM suggest that the systemic maternal inflammatory response during preterm PROM [8] may contribute to the increased median maternal plasma TF concentration in these patients regardless to the presence of infection or inflammation in the amniotic cavity or the occurrence of vaginal bleeding.

The interaction between the coagulation cascade and inflammation is well-established [64-66]. Patients with chronic and acute inflammation or infection (i.e. chronic and acute pancreatitis [67,68] non-crescentic glumerolonephritis [69] and meningococcal sepsis [70]) have higher plasma TF concentrations than healthy controls. During preterm PROM, there is a moderate maternal systemic inflammation that results in monocyte and granulocyte activation [8]. Activated monocytes express TF on their membrane [71-75] and shed micro-particles containing TF into the plasma [71]. In addition, the lack of association between intra-amniotic infection/inflammation, as well as the placental histologic findings and median maternal plasma concentrations of TF and TFPI, suggest that the procoagulant changes observed in patients with preterm PROM may be due to a systemic rather than a local (i.e. placental, intrauterine) inflammatory process.

The procoagulant activity of immunoreactive TF in the maternal plasma (blood-born TF) is a topic of debate [71,72],[76-79]. In a recent in vitro study, blood-born TF had very little or no procoagulant activity [71]. Moreover, whole blood and plasma clot formation, after inhibition of the contact factor (factor XIIa), was obtained only by adding exogenous, active TF. Of note, only six fentomol of exogenous TF were sufficient for clot formation [71]. On the other hand, it has been proposed that blood-born TF does not initiate the coagulation cascade but rather propagates clot formation by attaching to activated platelets and further enhancing the coagulation process [76-79]. This might be the reason why the administration of anti-TF antibodies has an anti-thrombotic effect without causing severe bleeding; because these antibodies inhibit blood-born TF at concentrations below those needed for inhibiting its hemostatic effect [80]. Moreover, positive immunostaining of arterial thrombosis and atherosclerotic plaques for TF suggests that blood-born TF may participate in the propagation of atherosclerotic plaques and the generation of arterial thrombosis [80,81]. Further, higher concentrations of blood-born TF were predictive of cardiovascular mortality of patients with acute coronary syndrome [82,83]. Therefore, the higher blood-born TF concentration in patients with preterm PROM, compared with those with normal pregnancies, may result from the association between moderate maternal inflammation and an increased activation of the coagulation cascade, whether or not intra-amniotic infection or inflammation occurs.

What is TFPI?

The main physiological inhibitor of the TF pathway of coagulation is TFPI. Two types of TFPI have been described: TFPI-1 and TFPI-2. TFPI-1 directly inhibits factor Xa through the Kunitz-2 domain [84], whereas the inhibition of the FVIIa/TF complex is performed by the Kunitz-1 domain in a factor Xa-dependent manner [20],[85,86]. TFPI-1 is found in the maternal circulation, fetal blood, platelets, endothelial cells and other organs [24,25] and it is the major form of TFPI in the plasma of non-pregnant individuals [87,88].

TFPI-2, first isolated from human placenta as PP5 [32,33], is the major form of TFPI in the placenta [26-31], and it is expressed by the syncytiotrophoblast [35,36]. In the non-pregnant state, TFPI-2 is present mainly in the extracellular matrix and its plasma concentrations are very low [33-35]. Recombinant TFPI-2 inhibits trypsin and factor VIIa activity (in a dose-dependant manner). High concentrations of TFPI-2 are needed for a weak inhibition of factor Xa [89]. Moreover, TFPI-2 does not affect thrombin activity [89]. The administration of heparin reduces the dose of TFPI-2 needed for factor VIIa inhibition, but it does not increase the inhibitory activity of TFPI-2 on factor Xa [89]. In addition to its anticoagulant activity, TFPI-2 has a localised inhibitory activity on serine protease (i.e. trypsin, plasmin, plasma kalikrein) [28],[89-93], as well as the turnover of pro-MMP1 and pro-MMP3 to their active forms [94].

Heparin and low-molecular-weight heparins increase the production and secretion of TFPI by the endothelial cells [95-100], leading to an increase in the plasma concentrations of both types of TFPI [95],[97-112]. Heparin binds factor Xa and TFPI simultaneously, bringing them into proximity that enhances factor Xa inhibition by TFPI [20],[85],[113,114].

What are the changes in TFPI during pregnancy?

During normal pregnancy, the maternal plasma concentration of TFPI-2 increases along gestation reaching a plateau at 36 weeks [33-37] and subsiding after delivery [38]. Twin pregnancies have higher median maternal plasma concentration of TFPI-2 than singleton pregnancies, but the rate of the increase in TFPI-2 during gestation is lower in twins than in singletons [37]. However, the mean maternal plasma concentrations of total TFPI have been reported to increase during the first half of pregnancy until 20 weeks of gestation and then remain relatively constant until term [22]. Conflicting data exists concerning the changes in maternal plasma concentrations during labour. Although some authors described an increase in TFPI plasma concentrations in early stages of labour [87], others reported a decrease in the TFPI plasma concentrations during labour [23].

Decreased median TFPI plasma concentrations and low TFPI/TF ratio in patients with preterm PROM: an additional mechanism for increased thrombin generation?

The procoagulable state reported in preterm PROM may be, in part, due to reduced natural anticoagulant concentrations and not only derived from the higher concentrations of procoagulant proteases (i.e. thrombin). This is supported by the novel finding of this study that patients with preterm PROM have a lower median total maternal plasma TFPI concentration than in women with normal pregnancies. Similar association between low plasma concentrations of TFPI and procoagulant state have been reported in non-pregnant and pregnant women.

In non-pregnant patients: (1) low TFPI concentrations are considered an independent risk factor for deep vein thrombosis [115]; (2) among patients undergoing IVF treatments, those who developed the procoagulable state of ovarian hyper-stimulation syndrome, have lower TFPI plasma concentrations than those who did not [116]; and (3) non-pregnant women with a history of recurrent pregnancy loss [117] and a higher rate of thrombophilic mutation, have lower plasma TFPI concentrations than non-pregnant healthy women.

During pregnancy: (1) pregnant women with vascular complications of pregnancy (pre-eclampsia, eclampsia, placental abruption, fetal growth restriction and fetal demise) have lower placental extract total TFPI concentrations and TFPI mRNA expression than those of women with normal pregnancies [43]; and (2) a different report demonstrated a lower placental immunoreactivity of TFPI-2 in patients with pre-eclampsia but not in patients with fetal growth restriction [118]. Collectively, our results suggest that lower concentrations of total TFPI may contribute to the maternal procoagulant state observed in patients with preterm PROM.

The finding of a significantly lower TFPI/TF ratio in patients with preterm PROM than in women with normal pregnancy is novel and represents the procoagulant state in the maternal plasma of patients with preterm PROM. Immunoreactive TFPI is 500 to 1000 times more abundant in the maternal plasma than TF [119]. Changes in this ratio have been observed in patients with disseminated intravascular coagulation (DIC) [119] and thrombotic thrombocytopenic purpura (TTP) [79]. Moreover, patients with DIC who had a poor outcome also had a higher TF/TFPI ratio [119], whereas patients with TTP had a significant increase in TFPI/TF ratio after treatment; the authors had proposed that this reflects an improvement in the hypercoagulable state associated with TTP [79]. Therefore, the ratio of TFPI/TF can serve as an additional marker for increased activation of coagulation cascade.

In conclusion, preterm PROM is associated with increased median maternal plasma TF and decreased TFPI concentrations. This may result from the moderate systemic inflammatory state associated with preterm PROM but also from activation of the coagulation cascade as reflected by the low TFPI/TF ratio. Moreover, we suggest that measurement of the ratio between TF and its inhibitor (TFPI) may be of benefit in the assessment of hypercoagulable states during pregnancy.

Acknowledgements

This research was supported by the Intramural Research Programme of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH, DHHS.