Pregnancy involves a state of physiological hypercoagulation that minimizes blood loss during delivery but also increases the risk of thromboembolism [1]. Hypercoagulability is caused by increased levels of several procoagulants, including factors VII, VIII, X, and fibrinogen, and decreased levels of natural coagulation inhibitors, protein C, and protein S [2]. Activated protein C regulates coagulation via the neutralizing factors Va and VIIIa. Protein S is a vitamin K-dependent plasma protein that serves as a cofactor for activated protein C [3]. In the plasma, 60% of protein S is bound to the C4b-binding protein, whereas the remaining proteins are in the free form. Only the free form shows cofactor activity [4].

Most hemostatic parameters differ between pregnant and nonpregnant women [5]. Racial differences also influence the concentration and activity of various hemostatic factors [6]. Therefore, well-known coagulation parameters based on Caucasian pregnant women are inappropriate for assessing other races, which may prevent an accurate diagnosis. Establishing ethnicity-specific reference ranges for hemostatic parameters and determining changes in activity during normal pregnancy is necessary for appropriate management.

Thrombophilia is also associated with adverse pregnancy outcomes and may increase the risk of placental vascular complications, recurrent pregnancy loss, preeclampsia (Pre-E), and fetal growth restriction (FGR) [7-10]. However, the association between protein C and protein S activity levels and adverse pregnancy outcomes in healthy pregnant women without thrombophilia has not been clearly established, and there are few studies on people of different races. Therefore, for the clinical application of protein C and protein S activity values in normal pregnancies, studies using more defined subjects and test methods are needed to specifically define activity values in Korean women.

This study was conducted to analyze trimester-specific changes in protein C and S activities in pregnant Korean women. We also evaluated whether changes in the activity of these anticoagulants are related to adverse pregnancy outcomes.

1. Participants and data collection

This prospective study included 19–45 years old singleton pregnant women and age-matched non-pregnant women who visited Chungnam National University Hospital between April 2019 and April 2020. All participants were from the Korean population. Pregnant women were recruited based on their number of gestational weeks: first trimester (from conception to 14 weeks), second trimester (15–28 weeks), third trimester (29–42 weeks), and postpartum period (within 6 weeks after delivery). Women were not eligible if they had a history of recurrent (≥2 times) or mid-trimester miscarriages, Pre-E, or placental abruption. Women were also excluded if they had a history of deep vein thrombosis or pulmonary embolism, hereditary or acquired thrombophilia, chronic hypertension, pre-gestational diabetes (pre-GDM), nephropathy, acute or chronic infection, autoimmune disease, or tumors. Patients who used any medication that could affect coagulation, such as coumarin, lowmolecular- weight heparin, or oral contraceptives, were also excluded. The enrolled pregnant women were followed up until delivery, and blood samples were collected during each trimester and in the postpartum period. Blood samples were collected from nonpregnant women at the time of study enrollment.

The baseline characteristics of pregnant women, such as age, body mass index, parity, maternal weight gain during pregnancy, hemoglobin difference after delivery, delivery mode, gestational age at delivery, and birth weight of newborns, were collected. The primary outcome was trimester-specific changes in protein C and protein S activities. In secondary outcome analysis, we analyzed the association between adverse pregnancy outcomes and protein C and protein S activities. Adverse pregnancy outcomes, such as preterm delivery, premature rupture of membranes, FGR (<10th percentile for gestational age), large for gestational age (>90th percentile for gestational age), intrauterine fetal death, pregnancy-induced hypertension (PIH)/Pre-E, GDM, deep vein thrombosis, and placental abruption were considered [11].

2. Measurement of protein C and protein S activities

Blood samples were collected through venipuncture of the antecubital vein into tubes containing 3.2% sodium citrate at a ratio of 9:1 (Greiner Bio-One GmbH, Kremsmünster, Austria). The samples were centrifuged, and the plasma was stored at -70°C until further testing. Protein C and protein S assays were performed using an automated coagulation analyzer (ACL TOP 550 CTS; Instrumentation Laboratory, Bedford, MA, USA) according to the manufacturer’s instructions. Protein C and protein S activities were measured using a chromogenic assay with a HemosIL Protein C kit and clotting time method HemosIL Protein S activity kit (Instrumentation Laboratory), respectively. The results of protein C and protein S activities were expressed in percent of normal plasma activity (activity in 1 ml of pooled plasma=100% activity).

3. Statistical analysis

The results were analyzed using SPSS Statistics for Windows, version 22.0 software (SPSS, Inc., Chicago, IL, USA). For distribution testing, we used the Shapiro-Wilk test. All parameters were reported as median (interquartile range) values. Protein C and protein S activities for coagulation assays are presented with their 95% confidence intervals (CIs), calculated as the mean±1.96×standard deviation. Kruskal-Wallis test was used to compare significant differences in the levels of protein C and protein S activities among groups. Mann-Whitney U test was used to examine the relationship between protein C and protein S activities and adverse pregnancy outcomes. Receiver operating characteristic (ROC) curve analysis was conducted to assess the predictive ability of protein C and protein S activities for adverse pregnancy outcomes. The area under the curve (AUC) values were calculated to determine the discriminatory power of protein C and protein S activities in identifying specific adverse outcomes. Linear logistic regression analysis was performed to determine the effect of abnormal protein activity on adverse pregnancy outcomes. All tests were two-sided, and P-value <0.05 were considered to indicate statistically significant results. 4. Ethics statement This study was approved by the Institutional Review Board of Chungnam National University (IRB No. 2019-04-041-003). Written informed consent was obtained from all participants.

A total of 225 Korean women (172 singleton pregnant women and 53 non-pregnant women) was enrolled in the study. The pregnant group was comprised of those who delivered at Chungnam National University Hospital. Of the 172 pregnant women, 21, 31, 66, and 54 were assigned to the first, second, third trimester, and postpartum groups, respectively (Fig. 1). The baseline characteristics of these women are presented in Table 1. There were no significant differences in age and body mass index between the pregnant and non-pregnant groups (P=0.181 and 0.487, respectively).

Figure 1. Flowchart for enrolled participants.

Table 1 Baseline characteristics of pregnancy and non-pregnancy groups

Characteristics	Pregnancy (N=172)	Non-pregnancy (N=53)	P-value
Age	33.5 (31.0–37.05)	38.0 (24.0–52.0)	0.181
Pre-pregnancy BMI	22.1 (20.2–24.21)	22.0 (17.2–26.8)	0.487
Pre-delivery BMI	27.0 (24.9–28.9)
Parity
Primigravida (N)	103 (59.9%)
Multigravida (N)	69 (41.1%)
Maternal weight gain during pregnancy (kg)	12.4 (8.2–17.6)
Hemoglobin difference after delivery (g/dL)	1.7 (1.2–2.3)
Delivery mode
Vaginal delivery (N)	83 (48.3%)
Cesarean section (N)	89 (51.7%)
GA at delivery (weeks)	38.3 (35.5–40.1)
Birth weight of newborn (kg)	3.0 (2.7–3.3)

Data are expressed as the median (interquartile range) or number (percentage).

Abbreviations: BMI, body mass index; GA, gestational age.

Table 2 shows the trimester-specific protein C and protein S activities in each group. Because of the insufficient sample volume, protein S but not protein C could be tested in 20 samples; thus, the sample numbers of protein C and protein S differed. The median protein C activity was 101.0%, 121.0%, 107.5%, 113.0%, and 78.1% in the first trimester, second trimester, third trimester, postpartum, and non-pregnant groups, respectively. The median protein S activities in the first trimester, second trimester, third trimester, postpartum, and non-pregnant groups were 36.0%, 39.9%, 41.6%, 63.6%, and 83.0%, respectively. Fig. 2 shows the differences in protein activity between groups. Protein C activity was significantly increased in the second trimester, third trimester, and postpartum periods compared with that in the non-pregnant group (P<0.001, P=0.017, and P<0.001, respectively). Protein S activity was significantly decreased from the first trimester to the postpartum period compared to that in the non-pregnant group (P<0.001, P<0.001, P<0.001, and P=0.011, respectively). Protein C and protein S activities were not associated with adverse pregnancy outcomes in the first and second trimesters or in the postpartum groups. However, protein C and protein S activities in the third trimester were correlated with the occurrence of adverse pregnancy outcomes. Table 3 shows the association between protein C and protein S activity in the third trimester and adverse pregnancy outcomes. There were no cases of intrauterine fetal death or placental abruption. Protein C activity was significantly higher in women who gave birth to babies with FGR than in women who had babies without FGR (40.1% vs. 27.6%, P=0.040). Women with PIH/Pre-E and GDM showed significantly lower protein S activity than women without these complications (10.1% vs. 35.0%, P=0.008; and 25.9% vs. 38.8%, P=0.042, respectively).

Figure 2. Relative protein C and protein S activities (%) during pregnancy compared to those in non-pregnant women. The protein C and protein S activities in non-pregnant women were set at 100%, and the relative value of protein C and protein S activities of each trimester and in the postpartum period were calculated. The blue line represents the relative value of protein C activity. The orange line represents the relative value of protein S activity. An asterisk (*) indicates a significant difference (*P<0.05).

Table 2 Medians and 95% confidence intervals of protein C and protein S activities by trimester in pregnant and non-pregnant women

Group	Protein C activity (%)			Protein S activity (%)
	N	Median (IQR)	95% CI	N	Median (IQR)	95% CI
Pregnancy	152	110.0 (83.0–130.8)	103.9–113.9	172	45.0 (30.8–62.9)	45.1–54.7
First trimester	21	101.0 (71.5–114.0)	84.5–107.4	21	36.0 (26.9–68.3)	33.1–57.2
Second trimester	24	121.0 (97.5–140.0)	105.4–130.5	31	39.9 (19.9–48.0)	30.5–48.1
Third trimester	58	107.5 (77.3–130.3)	97.6–114.3	66	41.6 (23.8–45.8)	31.5–41.4
Postpartum	49	113.0 (85.0–134.0)	104.3–122.7	54	63.6 (52.6–76.3)	61.6–84.0
Non-pregnancy	53	78.1 (67.6–104.7)	79.3–93.7	53	83.0 (58.7–137.2)	83.9–112.7

Data are expressed as the median (interquartile range), where N is the number of samples in each group.

Abbreviations: IQR, interquartile range; SD, standard deviation; 95% CI, 95% confidence interval (mean±1.96 SD).

Table 3 Comparison of differences in protein C and protein S activities in third trimester with and without adverse pregnancy outcomes

Pregnancy outcomes	Protein C activity (%)				Protein S activity (%)
	N (%)	Occur	Con	P-value	N (%)	Occur	Con	P-value
Preterm delivery	2 (3.4)	19.5	29.9	0.437	2 (3.0)	21.0	33.9	0.392
PROM	1 (1.7)	27.0	29.5	0.931	1 (1.5)	25.0	33.6	0.758
FGR	9 (15.5)	40.1	27.6	0.040*	9 (13.6)	41.1	32.3	0.204
LGA	1 (1.7)	14.0	29.8	0.483	1 (1.5)	7.0	34.0	0.212
IUFD	0				0
PIH/ Pre-E	4 (6.9)	18.3	30.3	0.179	4 (6.1)	10.1	35.0	0.008*
GDM	11 (19.0)	30.6	29.2	0.812	12 (18.2)	25.9	38.8	0.042*
DVT	1 (1.7)	42.5	29.3	0.552	1 (1.5)	42.0	33.4	0.758
Placenta abruption	0				0
PPH	1 (1.7)	21.0	29.7	0.724	2 (3.0)	46.0	33.1	0.392

Data are expressed as the median or number (percentage).

*P<0.05.

Abbreviations: Occur, occurrence group; Con, control group; PROM, premature rupture of membrane; FGR, fetal growth restriction; LGA, large for gestational age; IUFD, intrauterine fetal death; PIH, pregnancy-induced hypertension; Pre-E, preeclampsia; GDM, gestational diabetes mellites; DVT, deep vein thrombosis; PPH, postpartum hemorrhage.

For factors showing significant differences (Table 3), an ROC curve was used to determine protein activity, which is a meaningful cutoff for predicting adverse pregnancy outcomes (Fig. 3). The AUC of protein C activity, which was associated with an increased risk of FGR, was 0.717 (95% CI 0.517–0.916, P=0.040) (Fig. 3A), with a cutoff value of 113.0% (sensitivity, 0.667 and specificity, 0.633). The AUC of protein S activity that increased the risk of PIH was 0.877 (95% CI 0.775–0.979, P=0.012), and the cutoff value was 25.8% (sensitivity, 0.774 and specificity, 0.750) (Fig. 3B). Finally, the AUC of protein S activity that increased the risk of GDM was 0.661 (95% CI 0.472–0.851, P=0.082), with a cutoff value of 35.9% (sensitivity, 0.630 and specificity, 0.667) (Fig. 3C). The women were divided into groups based on the obtained cutoff values, and event occurrence probabilities were compared (Table 4). According to the cutoff criteria for FGR, PIH/Pre-E, and GDM, only PIH/Pre-E showed a significant increase in incidence when protein S activity was less than 25.8% (odds ratio, 0.156; P=0.020).

Figure 3. Predictive value of protein C and protein S activities in the third trimester of pregnancy for adverse pregnancy outcomes with receiver operating characteristic (ROC) curve. (A) The area under the curve (AUC) of protein C activity associated with the risk of fetal growth restriction (FGR) was 0.717 (95% CI 0.517–0.916, P=0.040), and the cutoff value was 113.0 (sensitivity 0.667, specificity 0.633). (B) AUC of protein S activity associated with the risk of pregnancy-induced hypertension (PIH) was 0.877 (95% CI 0.775–0.979, P=0.012), and the cutoff value was 25.8% (sensitivity 0.774, specificity 0.750). (C) AUC of protein S activity associated with the risk of gestational diabetes (GDM) was 0.661 (95% CI 0.472–0.851, P=0.082), and the cutoff value was 35.9% (sensitivity 0.630, specificity 0.667).

Table 4 Abnormal protein C and protein S activities and adverse pregnancy outcomes examined using linear logistic analysis

Pregnancy outcomes	Protein C activity above 113.0% (N=24)		Protein S activity below 25.8% (N=17)		Protein S activity below 35.9% (N=28)
	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value
FGR	0.162 (-0.030–0.354)	0.097
PIH/Pre-E			0.156 (0.026–0.287)	0.020*
GDM	0.180				(-0.009–0.370)	0.062

*P<0.05.

Abbreviations: OR, odds ratio; 95% CI, 95% confidence interval (mean±1.96 SD); FGR, fetal growth restriction; PIH, pregnancy-induced hypertension; Pre-E, preeclampsia; GDM, gestational diabetes mellitus.

We analyzed the differences in protein C and protein S activities during each trimester and postpartum period in pregnant Korean women. Protein C activity tended to increase starting in the first trimester compared to that in non-pregnant women, with a significant increase from the second trimester to the postpartum period (Table 2, Fig. 2). This finding is consistent with those of previous studies [12, 13]. Protein S activity was reduced throughout pregnancy and in the postpartum period, with a more than 50% decrease in the first trimester compared to that in non-pregnant women, followed by a minimal increase. This result is consistent with those of previous studies showing a decline in protein S activity during pregnancy. Studies predominantly in Asian women showed a dramatic decrease in protein S activity from the first trimester, with no further decrease before delivery [14, 15]. However, several studies of Western pregnant women reported a decrease in protein S activity, particularly in the third trimester [16, 17]. Potential explanations for racial differences in the coagulation mechanisms include genetic mutations, different levels of coagulant and anticoagulant proteins, and altered anticoagulant drug metabolism [18]. Describing normal plasma values of coagulation factors according to race is important for providing appropriate criteria when selecting thromboprophylaxis during pregnancy and reducing the risk of pregnancy complications.

Venous thromboembolism is a pregnancy complication and leading cause of maternal death, accounting for 1.1 maternal deaths per 100,000 live births [3]. There are racial differences in the prevalence and risk factors for the development of this condition. Factor V or Factor II polymorphisms are major genetic risk factors in Western ethnic groups but are rare in Asians [19, 20]. In contrast, congenital deficiency of protein S has been reported more frequently in Asians than in Caucasians [21, 22]. We observed significant changes in protein C and protein S activities during the pregnancy. Particularly, protein S activity was decreased from the first trimester, which may be important in the development of thrombotic events in pregnant Korean women. One participant in our study presented with left leg swelling at week 14 of gestation and left femoral vein thrombosis was diagnosed using ultrasonography. She had hyperemesis and lost 15% of her body weight before pregnancy. Her blood test showed a protein S activity of 8%, which was significantly lower than the normal value in our study (10th percentile, 14%). Her protein C activity was 79%, which was within the normal range. This case suggests that excessively decreased protein S activity in the first trimester contributes to thrombotic events during pregnancy.

Hypercoagulability is observed during pregnancy; however, most patients maintain normal pregnancy without complications. Moreover, when excessive changes in coagulation or fibrinolysis occur beyond the normal range, adverse pregnancy outcomes can occur. We found that an excessive increase in protein C activity was associated with FGR and that an excessive decrease in protein S activity was associated with PIH/Pre-E and GDM (Table 3). In agreement with these findings, previous studies reported that protein C and protein S deficiencies are significantly increased in FGR [10], and Pre-E is more likely to result in protein S deficiency [7]. According to a study conducted in Japan, women with low plasma protein S activity during early pregnancy may be at an increased risk of PIH and Pre-E [9]. Hypercoagulability is exacerbated in pregnant women with GDM, and low levels of protein S activity are associated with glucose metabolism and inversely related to glycated hemoglobin levels [23, 24].

Our results show that abnormal changes in Protein C and S activities increase the risk of developing FGR and PIH/Pre-E (Fig. 3A and B). Placental dysfunction is the leading cause of both Pre-E and FGR [25]. Recent evidence suggested that the protein C anticoagulant system regulates the growth and survival of trophoblast cells in the placenta, thereby preventing utero-placental circulation from local thrombosis in decidual vessels [26, 27]. Protein S directly activates C-independent anticoagulant activity by inhibiting the formation of prothrombin-proteinase complexes [28]. Thus, excessive and abnormal protein activities may induce thrombosis in decidual vessels and impair placentation through hypercoagulability and inflammation, leading to adverse pregnancy outcomes [29].

The main limitation of our study was the small number of patients evaluated. The initial purpose of our study was to determine the reference interval of protein activity in healthy pregnant Korean women. To present the reference interval, at least 120 samples from women in each trimester should be used based on the CLSI guideline EP28-A3c [30]; thus, we could not evaluate the reference interval. Although the pattern of protein activity changes in pregnant Korean women was confirmed, further studies are needed to determine the reference interval value using a larger sample size. In addition, our statistical results may have been influenced by the small number of mothers with adverse pregnancy outcomes. Furthermore, we assessed only the activity of proteins C and protein S in functional assays and did not examine other parameters, such as total or free antigen or C4b-binding proteins. Further studies of a larger number of participants and involving various functional assessment parameters are required to support our findings. However, our prospective study revealed changes in protein C and protein S activity in pregnant women in Korea and the relationship between protein C and protein S activity and adverse pregnancy outcomes.

In conclusion, we present trimester-specific changes in the activities of proteins C and protein S in normal pregnant Korean women. Protein C activity significantly increased from the second trimester of pregnancy to the postpartum period, and protein S activity significantly decreased throughout pregnancy and in the postpartum period compared with those in non-pregnant women. Protein S activity showed a particularly sharp decrease during the first trimester. In the third trimester, an increase in protein C activity was associated with FGR, and a decrease in protein S activity was associated with PIH/Pre-E and GDM. PIH/Pre-E rates were significantly increased when protein S activity was below the cutoff value of 25.8%. Further studies are needed to establish an reference interval of coagulation factors for Korean pregnant women that can be used to predict adverse pregnancy outcomes.

This study was supported by Chungnam National University Hospital Research Fund (2016).

None declared.