Prior Preterm Preeclampsia and Hemostatic Balance in the Non-pregnant State

Abstract

Objective

To evaluate whether a history of preterm preeclampsia (PreP) is associated with abnormal hemostatic features during the non-pregnant state, relative to people without a history of preeclampsia.

Study Design

We conducted a retrospective sub-analysis of three prospective IRB approved studies in which individuals were evaluated outside of pregnancy during the follicular phase of their menstrual cycles. We compared parameters between subjects with prior PreP (n=21) with a combined cohort (NL, n=74) of healthy nulliparous subjects (n=57) and subjects with a prior normotensive pregnancy (n=17). T-tests were employed with P < 0.05 accepted for significance.

Main Outcome Measures

Levels of eight coagulation factors, protein C (PC), and thrombin generation (TG) were assessed in the presence or absence of thrombomodulin (TM) to evaluate the activation of the anticoagulant PC pathway. Results: There were no differences in TG or factor composition between healthy nulliparous and healthy primiparous people. In contrast, the PreP cohort had increased levels of factors II, V, VII, IX, X, XI, and PC compared to the NL cohort without this history. In dynamic assays examining TG in the presence of TM, the PreP group had reduced rates (nM/min) of formation and diminished peak levels when compared to the NL group.

Conclusions

Increased procoagulant factor levels are typically associated with increased TG, suggesting here that individuals with prior PreP would have a procoagulant shift in the non-pregnant state. However, the observed concurrent increase in anticoagulant PC linked with a decrease in dynamic TG suggests a possible compensatory mechanism to maintain hemostatic balance.

Introduction

Pregnancy is clinically recognized as a pro-thrombotic state, and it is hypothesized that changes in coagulation factors during pregnancy are protective, given the significant risk of obstetric hemorrhage. During normal pregnancy, concentrations of many coagulation factors increase, including factors XIII, XII, X, VIII, VII, and fibrinogen [1]. Protein C (PC) levels typically remain unchanged during normal pregnancy [2,3].

Despite the potential physiologic contribution of the hypercoagulable state during pregnancy and the early postpartum period, adverse thrombotic events also occur much more frequently during pregnancy and the postpartum period compared to age-matched, non-pregnant individuals. The incidence of venous thromboembolism (VTE) has been shown to range from 0.49 to 1.72 per 1000 pregnancies [4]. VTE risk is estimated to be approximately 4x greater in pregnant and peripartum individuals than in non-pregnant people of the same age [5], with the risk of thrombosis at its highest during the postpartum period, defined in this study as up to 3 months following delivery [5]. A meta-analysis reported the estimated relative distribution of 100 deep vein thrombosis events during pregnancy and the postpartum period would be 0.23/day during pregnancy and 0.82/day during the postpartum period, with most included studies defining the postpartum as up to six weeks following delivery [6]. Pregnancy-associated changes in coagulation profiles have been shown to return to their pre-pregnancy status by approximately 12 weeks postpartum in most individuals [1], though the rates at which specific coagulation factors and other hemostatic proteins return to a pre-pregnancy baseline vary [7, 8]. There are several conditions of pregnancy that have been shown to be correlated with aberration of this normal hemostatic balance, such as preeclampsia.

Preeclampsia is a complication of pregnancy that presents after 20 weeks gestation, with high blood pressure and evidence of end organ injury [9]. Preeclampsia affects approximately 5–7% of all pregnancies and is associated with significant neonatal and maternal morbidity and mortality [10, 11, 12]. While the mechanism(s) underlying the development of preeclampsia remain unknown, there is substantial evidence demonstrating that a history of preeclampsia is associated with increased future maternal cardiovascular risk, including thrombotic events [10, 11, 13].

Preeclampsia appears to be a condition in which the general pro-thrombotic state of pregnancy is exaggerated [14]. The mechanism(s) behind this enhancement are unclear, but may be multifactorial and involve endothelial cells, platelets, fibrinolysis, and the coagulation cascade. Feldstein et al. evaluated the PT, aPTT, INR, and fibrinogen of 348 individuals with a preeclamptic pregnancy and found that 16.1% had abnormal coagulation profiles, evaluated at an average of 4.2 ± 1.6 hours prior to delivery [15]. Further, individuals with preeclampsia have been shown to have an increased rate of thrombin generation (TG) [16], more activated platelets [17], and more procoagulant placenta-derived extracellular vesicles [18] in the third trimester than normotensive pregnant people. The difference in coagulability is noted clinically, as those with preeclampsia may be up to 5 times more likely to develop VTE relative to their normotensive counterparts [19].

Despite the reported changes in coagulability in individuals with preeclampsia during pregnancy and the early postpartum period, the duration, nature, and contributing mechanisms of the abnormal coagulation profiles beyond the early postpartum period is not well-understood. Here, we assess whether a history of preterm preeclampsia is associated with abnormal hemostatic features during the non-pregnant state, remote from pregnancy and the postpartum period, relative to those without a history of preterm preeclampsia.

Methods

Study participants:

This was a retrospective analysis of individuals who enrolled in one of three different human subjects studies approved by the University of Vermont Institutional Review Board. Written consent was obtained from all participants prior to enrollment. All individuals were not pregnant at time of enrollment. Subjects were between the ages of 18 and 44. Enrollment groups targeted those trying to conceive and were either: nulliparous, parous without a personal history of a preeclamptic pregnancy, or parous with a personal history of a preterm preeclamptic pregnancy. Exclusion criteria included active malignancy, infection, systemic inflammatory conditions, type 2 diabetes mellitus, smoking, 1st degree relative with thrombosis, known thrombophilia, and collagen vascular disease. Patient demographic characteristics including age, body mass index (BMI), race, and menstrual cycle day at time of study visit were collected. Obstetric and fetal delivery data were collected via chart review. Two participants with prior preterm preeclampsia had chronic hypertension and were taking daily blood pressure medications at the time of study enrollment.

Visits and Sample Collection:

All study visits were performed at the University of Vermont Medical Center’s Clinical Research Center. All study visits occurred during the follicular phase. Intravenous (IV) catheters were placed with subjects in a seated position. Blood was drawn once >15 minutes had elapsed from IV placement, collected into syringes, and immediately distributed to 3.2% sodium citrate (9:1 blood:citrate ratio). Citrated plasma was centrifuged at 3400 g for 17 min for two of the included studies and 1865 g for 10 minutes in the third study, and the resulting platelet poor plasma aliquoted and stored at −80°C for subsequent analysis.

Thrombin Generations (TG) Assays:

Plasma aliquots were thawed at 37°C for 5 minutes in the presence of 0.1 mg/mL corn trypsin inhibitor to prevent activation of the contact pathway of coagulation. TG was measured in a Stago Thrombinoscope using the established Calibrated Automated Thrombogram protocol and standard reagents. Briefly, thawed plasma was pre-incubated for 10 minutes at 37°C with an initiation reagent containing small amounts of phospholipids with exogenous tissue factor (TF, Stago’s commercial platelet rich plasma reagent). Reactions were triggered by the addition of the FluCa reagent containing CaCl₂ and a fluorescent thrombin substrate used to monitor TG over time. TG curves were characterized using various parameters: 1) lag time; 2) peak thrombin level; 3) maximum rate, and 4) endogenous thrombin potential (ETP). Additional assays were performed in the presence of 10 nM (relative to plasma) soluble human recombinant thrombomodulin (TM, Cambridge Protein Works, Cambridge, UK) to initiate the inhibitory protein C (PC) pathway (TF + TM). TG assays were performed in duplicate; TG assays with technical errors were excluded in the analysis.

Coagulation Factor Level Measurements:

Plasma activity concentrations of factors II, V, VII, VIII, IX, X, XI, XII, and protein C (PC) were determined by clot based assays at either the CLIA certified University of Vermont Medical Center Coagulation Laboratory (Burlington, VT) or the Laboratory for Clinical Biochemistry Research at the University of Vermont (Colchester, VT). Factor levels were not measured in 4 PreP and 6 NL subjects due to insufficient plasma volume. Plasma antigen concentrations of tissue factor pathway inhibitor (TFPI) were measured using commercially available ELISA kits (TFPI_free : R&D Systems, Minneapolis, MN; TFPI_total: Stago, Parsippany, NJ) in the PreP and nulliparous individuals only.

Statistics:

Group comparisons on demographics, coagulation factor levels, and TG measures were done using t-tests for continuous variables and chi-square tests for categorical variables.

Results

Healthy nulliparous individuals (n=57), and individuals with a prior normotensive pregnancy (n=17) were compared. No differences between these two groups were observed (Supplemental Table 1); these cohorts were therefore used as a combined control group (NL, n=74) and compared against the preterm preeclampsia group (PreP) (Table 1). Study subjects with prior PreP were on average 2.5 years out from their prior preeclamptic pregnancy. The average mean arterial pressure for the overall study population was 85.18 ± 9.0 mmHg. We observed no significant difference between the PreP cohort and the NL cohort (PreP = 83.74 ± 8.7 mmHg; NL = 85.20 ± 9.0 mmHg, p=0.84). The average platelet count for the entire study population was within normal limits (215 ± 45 k/cmm). There was no difference in platelet count between groups (PreP = 221 ± 46 k/cmm, NL = 215 ± 50 K/cmm, p=0.56).

Table 1.

Demographic characteristics of study subjects by prior preeclamptic pregnancy status

Characteristic Avg ± S.D. or n (%)	PreP Group (n=21)	NL Group (n=74)	P-value
Age, years	30 ± 4	32 ± 4	0.08
Race/Ethnicity			.68*
White/Caucasian	20 (95)	67 (91)
Asian	0	5 (7)
Hispanic/Latina	1 (5)	1 (1)
American Indian or Alaskan Native	0	1 (1)
BMI	28 ± 7	27 ± 7	0.30
BMI by Category			.17
Underweight (<18.5)	0	1 (1)
Normal weight (18.5–24.9)	9 (43)	38 (52)
Overweight (25.0–29.9)	2 (9)	17 (23)
Obese (>30.0)	10 (48)	18 (24)
Menstrual cycle day at time of visit	10 ± 4	9 ± 4	.28

^*Comparison of white vs all other groups combined

Subjects with a history of prior PreP had significantly increased levels of factors II, V, VII, IX, X, XI, and PC compared to those without this history (Table 2). Factors VIII and XII did not differ between groups. In dynamic assays examining TG in the presence of TF alone, an assessment of TG without activators of the inhibitory protein C pathway presents the maximum TG generation potential, the PreP and NL groups displayed similar TG— no differences were seen in peak thrombin, max rate, or ETP (Table 3). However, PreP subjects had longer lag times compared to their counterparts without a history of preeclampsia (Table 3). In contrast, when assessing TG in the presence of TM, a modulator of coagulation and co-activator of PC, PreP individuals had decreased TG potential compared to NL individuals. Peak thrombin, max rate, and ETP were significantly lower; there was no difference in lag time (Table 4). In an analysis of tissue factor pathway inhibitor (TFPI), free TFPI was increased in the PreP group compared to healthy nulliparous women, while total TFPI did not differ (Supplemental Table 2). TFPI data for the cohort of subjects with normal prior pregnancies is not available.

Table 2.

Coagulation factor levels in study subjects with a history of preterm preeclampsia and controls

Coagulation Factor, % of Mean Population Avg ± S.D.	PreP Group (n=17)	NL Group (n=68)	P-value
Pro-coagulant
II	103.0 ± 13.8	91.5 ± 12.8	.002
V	97.4 ± 17.4	83.6 ± 14.6	.001
VII	112.0 ± 32.8	95.9 ± 24.7	.03
VIII	92.3 ± 23.0	99.6 ± 31.2	0.37
IX	119.9 ± 23.0	91.7 ± 19.7	<.001
X	102.7 ± 17.7	93.3 ± 14.5	.03
XI	116.8 ± 28.0	91.8 ± 25.8	<.001
XII	113.3 ± 25.1	103.7 ± 26.0	0.17
Anticoagulants
PC	142.4 ± 32.2	112.0 ± 20.7	<.001

Table 3.

In vitro measurements of thrombin generation initiated with tissue factor (TF) by cohort.

Thrombin Generation Parameters Avg ± S.D.	PreP Group (n=21)	NL Group (n=69)	p-value
Lag Time, min	3.0 ± 0.5	2.7 ± 0.5	.02
Peak Thrombin, nM	243.3 ± 55.3	237.2 ± 56.9	.66
Max Rate, nM/min	69.9 ± 27.0	77.0 ± 28.5	.31
ETP, nM•min	1567.2 ± 368.8	1411.7 ± 309.3	.06

Table 4.

In vitro measurements of thrombin generation initiated with tissue factor (TF) in the presence of thrombomodulin (TM) by cohort.

Thrombin Generation Parameters Avg ± S.D.	PreP Group (n=21)	NL Group (n=74)	p-value
Lag Time, min	2.6 ± 0.4	2.5 ± 0.5	0.51
Peak Thrombin, nM	100.1 ± 61.4	128.9 ± 49.9	0.03
Max Rate, nM/min	41.3 ± 27.2	53.8 ± 22.1	0.04
ETP, nM•min	471.4 ± 293.4	625.8 ± 257.4	0.02

The obstetric and fetal data collected via chart review did not demonstrate significant differences between groups other than gestational age at delivery and birth weight, which is not unexpected. Thirty-eight women had a pregnancy prior to this study. Of the 37 women with available data, 13 delivered via cesarean section (PreP cohort, n=11; normal prior pregnancy cohort, n=2) and 24 delivered vaginally (PreP cohort, n=10; normal prior pregnancy cohort, n=14). The gestational age at delivery for the PreP cohort was 33.3 ± 4.6 weeks and 40.0 ± 2.1 weeks for the normal prior pregnancy cohort. Eight of our participants used assisted reproductive technology, 3 with prior preeclampsia and 5 with normal prior pregnancies.

Of the 36 participants where neonatal sex was retrievable, 21 participants had female children, including one female/female twin pair (PreP cohort, n=12; normal prior pregnancy cohort, n=9), and there were 15 males (PreP cohort, n=8; normal prior pregnancy cohort, n=7). Newborn weight data from 34 participants was available (PreP cohort, n=20; normal prior pregnancy cohort, n=14). The PreP cohort had a lower mean newborn delivery weight than the normal prior pregnancy cohort (PreP cohort= 2069 ± 1065 grams, normal prior pregnancy cohort= 3344 ± 524 grams, p=.0003).

Discussion

In this study, we demonstrate that individuals an average of 2.5 years removed from a prior pregnancy complicated by PreP have different hemostatic profiles relative to a control group comprised of healthy nulliparous individuals and individuals with prior normotensive pregnancies.

Here, we observed that non-pregnant people years following a pregnancy complicated by PreP had increased procoagulant factor levels II, V, VII, IX, X, and XI relative to controls. Classically, increased procoagulant factor levels are associated with increased TG suggesting here that individuals with prior PreP would have a pro-coagulant phenotype outside of pregnancy. However, in our analysis of TG, we did not observe increased TG in the PreP group and instead observed that PreP and NL people had largely the same maximum TG potential when assessed with TF alone. When TG was assessed in the presence of TM, a co-activator of PC and thus modulator/regulator of thrombin generation, PreP individuals had decreased TG compared to their NL counterparts. Additionally, we analyzed TFPI levels and found that, compared to healthy nulliparous individuals only, individuals with prior PreP had similar levels of total TFPI but increased free TFPI, suggesting a mechanism of delaying and downregulating TG in PreP individuals. In summary, despite increases in pro-coagulant factors, the observed decrease in dynamic TG in the presence of TM coupled with an increase in PC levels in PreP individuals suggests a compensatory mechanism to maintain hemostatic balance. Clinically, despite known aberration in coagulation balance and the increased frequency of VTEs in people with preeclampsia, the absolute risk of pregnancy-associated VTE is not dramatically higher than in people without preeclampsia. A 2024 study that followed 522,545 primiparous participants with normotensive pregnancies and pregnancies complicated by preterm and term preeclampsia found that, though the participants with preeclampsia had higher incidences of VTEs during pregnancy and in the puerperium, the most dramatic difference in incidence of VTEs was seen at approximately 10 years following pregnancy [20]. Of note, participants with preterm preeclampsia had more VTEs than participants with late onset preeclampsia.

It is well reported that people with preeclampsia have different hemostatic profiles during pregnancy compared to people with normal pregnancies [1, 15, 16], and it is well-described clinically that those with early-onset, severe preeclampsia are at increased risk of VTE. A 2024 meta-analysis including 5,964 subjects found that people with preeclampsia (n=2,883) had significant increases in PT and aPTT compared to normotensive pregnant people (n=3,081), indicating an overall decrease in coagulation potential [21]. Of note, this meta-analysis includes study subjects with preterm and term preeclampsia and does not differentiate outcomes based on gestational age at diagnosis. However, other studies have shown individuals with preeclampsia to be hypercoagulable compared to normotensive pregnant individuals. A 2023 case-control study (n=180) found that those with preeclampsia, no matter time of onset, had shortened aPTT, higher D-dimer levels, and lower PC, protein S, and antithrombin levels compared to normotensive pregnant people [22]. Another study found increases in plasma levels of factors VII, VIIa, and TFPI in individuals with preeclampsia compared with healthy nonpregnant or healthy pregnant individuals [23].

Interestingly, the aberration in hemostatic balance during pregnancy appears to be correlated with severity of preeclampsia. A recent retrospective review of over 2.2 million subjects, 56,446 of whom had preterm preeclampsia, found that preterm preeclampsia was independently associated with an increased risk of thromboembolic events, while term preeclampsia did not incur additional risk [24]. The clinical thromboembolic outcome evidence reported by this study is supported by smaller, translational studies evaluating coagulation profiles. A 2007 study found that individuals with severe preeclampsia before 34 weeks had increased levels of factor VIIIR:Ag, fibrinogen, D-dimer, and TM, and decreased platelet counts compared to those who developed late preeclampsia [25]. A more recent study comparing people with early (n=26) and late (n=19) onset severe preeclampsia found increased levels of factorVIIa-antithrombin complex in those with severe preeclampsia, indicating an exaggerated procoagulant state in preeclampsia before 34 weeks [26]. Observed differences in the coagulation profile between those with preeclampsia versus those without have also been demonstrated in the postpartum period, though there are fewer studies specific to the puerperium. A 2021 study comparing 290 patients with preeclampsia against 256 without found that on postpartum day 1, the hemostatic balance of patients with preeclampsia is relatively low-coagulant and high-fibrinolytic, increasing the risk of postpartum hemorrhage [27].

The literature reports that normal pregnancy-associated changes in coagulation profiles return to their pre-pregnancy status at approximately 12 weeks postpartum in most individuals (Hellgren). This correlates clinically, as a 2012 study that included 1,687,930 participants found that there is an elevated risk of VTE up to 12 weeks postpartum compared to age and sex matched controls [28]. However, there is a paucity of data regarding when or if people with preeclampsia, who are known to have abnormal coagulation during pregnancy, return to a “normal” hemostatic balance in the postpartum period, and whether there is an affect associated with the gestational age at onset of preeclampsia during pregnancy.

Individuals with preeclampsia have increased risk for cardiovascular disease later in life compared to individuals with uncomplicated pregnancies [9, 29, 30]. This increased cardiovascular risk includes a persistent increased risk of VTE. A 2020 cohort study following 1,919,918 subjects, of which 26,280 had preeclampsia, found that a history of preeclampsia was associated with a significantly increased risk of thromboembolism up to 13 years following pregnancy, even after excluding thrombotic events during pregnancy and in the 3 months postpartum [31]. Perhaps over time, the compensatory adaptation we observed in this study population diminishes, leading to increased coagulability and susceptibility for adverse thrombotic events. Alternatively, this study could support the notion that individuals with preeclampsia are hemostatically different at baseline from people with uncomplicated pregnancies. The increased incidence of VTEs and CVD in people with histories of preeclampsia could be due to the compounding effects of age with a vulnerable phenotype at baseline.

Strengths of this study include the unique timing of primiparous participants’ evaluations, years after their pregnancies. To our knowledge, this is the first study to assess coagulation profiles in individuals with a history of PreP long after delivery. Our control group, consisting of both nulliparous individuals and those with previous normotensive pregnancies, which we found to be not statistically different, minimizes confounding factors, enabling a clearer assessment of the effects specific to preeclampsia. Furthermore, our comprehensive analysis of the coagulation cascade, including the PC pathway and TG, combined with our specific focus on individuals with a history of PreP, provides a robust and detailed evaluation of hemostatic changes.

However, this study has several limitations. The small sample size, limited ethnic and racial diversity of participants, and the absence of clinical outcome data are notable constraints. Additionally, the absence of TFPI data on participants with normal prior pregnancies limits our interpretations of TFPI data in conjunction with other data presented here. Further research is needed to explore hemostatic changes in individuals with preeclampsia across their lifespan, which will be crucial for developing targeted therapies and recommendations to reduce cardiovascular events in this high-risk population.

Highlights.

1. Preterm preeclampsia is associated with abnormal hemostasis years after pregnancy.

2. Coagulation factors and PC levels are increased; thrombin generation is decreased.

3. PC increases may compensate for increased coagulation factor levels.