Platelet reactivity changes significantly throughout all trimesters of pregnancy compared with the nonpregnant state: a prospective study

Authors


Abstract

Objective

Platelets play an important role in the pathophysiology of uteroplacental disease and platelet reactivity may be an important marker of uteroplacental disease activity. However, platelet reactivity has not been evaluated comprehensively in normal pregnancy. We sought to evaluate platelet reactivity using a number of agonists at defined time points in pregnancy using a novel platelet assay and compare these with a nonpregnant cohort.

Design

Prospective longitudinal study.

Setting

Outpatient department of a large tertiary referral centre.

Sample

Eighty participants with 30 nonpregnant women and 50 pregnant women assessed longitudinally.

Methods

This was a prospective cohort study performed longitudinally throughout uncomplicated singleton pregnancies with participants recruited before 15 weeks of gestation. They were controlled for a number of factors known to affect platelet reactivity. Blood samples were obtained in each trimester. Thirty nonpregnant healthy female volunteers also had a platelet assay performed. A modification of standard light transmission aggregometry was used to assess platelet function, with light absorbance measured following the addition of five different agonists at submaximal concentrations. Dose–response curves were plotted for each agonist for the nonpregnant cohort and in each trimester for the pregnant cohort.

Main outcome measures

Dose–response curves and median effective concentration.

Results

When compared with the nonpregnant controls a significant reduction was demonstrated in platelet reactivity to collagen during the first trimester of pregnancy (< 0.0001). Platelet aggregation increased significantly from the first to third trimesters in response to collagen and arachidonic acid.

Conclusion

Platelet reactivity varies according to pregnancy state, gestational age and agonist. The finding that platelet reactivity is reduced in the first trimester of pregnancy may be useful for the interpretation of further studies examining the role of platelet reactivity in the first trimester of pregnancies that develop uteroplacental disease.

Introduction

Platelets play a critical role in haemostasis and thrombosis. Although platelets have been implicated in a multitude of obstetric complications, their functional role has not been properly characterised in normal pregnancy. The ex vivo measurement of platelet responses to various agonists provides an index of the functional capacity of platelets, termed platelet reactivity. The laboratory investigation of platelet function usually involves measuring platelet aggregation. However, traditional platelet function tests have limited clinical application as they require blood sample processing, are time consuming, labour intensive and require skilled laboratory staff to perform and interpret the assay. Understanding the mechanisms of platelet aggregation in normal pregnancy, especially in early pregnancy, may provide useful information on the role of platelets in early pregnancy development, placentation and uteroplacental disease.

Platelet aggregation is a complex process involving multiple receptors responding to shearing forces, adhesion molecules and a plethora of agonists.[1] Platelets express several receptors, each with a distinct functional role, and it would seem logical to assess multiple receptors using several agonists to provide a more robust assessment of platelet reactivity. The correct method of assessing the mechanism of any receptor is by a dose–response stimulus or inhibition.

There are other factors that can influence platelet reactivity and result in inter-participant and intra-participant variability. Lifestyle factors such as smoking, circadian rhythm, eating, alcohol and exercise have all been shown to affect platelet reactivity.[2] Previous studies have not controlled for many of these factors. The maternal cardiovascular and haematological changes that take place from the first trimester through to the postpartum period are well described and may have substantial effects on platelet reactivity. Previous researchers have often only examined at a single time point in pregnancy. Longitudinal data are limited to much older techniques of platelet function with small sample sizes.[3]

With this in mind the objective of this study was to critically evaluate platelet reactivity, longitudinally, in normal pregnancy using a novel platelet assay and to compare these results with a healthy nonpregnant cohort. Strict selection and enrolment criteria were used and all of the participants had fasting samples taken. The unique platelet assay used in this study allowed us to measure responses to multiple agonists and therefore critique the various surface receptors on platelets expressed during pregnancy. Allowing for the adaptive physiological changes of pregnancy, we obtained blood samples in each trimester to give a comprehensive evaluation of platelet reactivity in pregnancy.

Methods

Ethical approval

Ethical approval was obtained from the Rotunda Hospital Research Ethics Committee before commencing this study and the study complied with the declaration of Helsinki. Informed consent was obtained from each participant before phlebotomy.

Study participants

Thirty healthy (nonpregnant) volunteers were recruited from the Rotunda Hospital. All of these participants were women with a history of a normal pregnancy outcome and no medical complaints. They were controlled for progesterone level (assay performed in the follicular phase) and all were nonsmokers. Participants were excluded if there was any history of nonsteroidal anti-inflammatory drug or aspirin use within the preceding 7 days, known thrombocytopenia, spherocytosis or coagulopathy, history of hereditary bleeding disorder, pre-eclampsia, recurrent miscarriage, or cardiovascular or inflammatory illness. Participants were asked to fast for at least 6 hours and avoid alcohol, coffee and strenuous exercise in the 24 hours before phlebotomy.

Fifty pregnant participants were recruited from the initial antenatal clinic visit in the Rotunda Hospital. They were telephoned in advance of their scheduled appointment and asked to fast if agreeable to participation in the study. Selection criteria included nonsmokers with a body mass index ≤ 30 kg/m2 with a singleton pregnancy ≤14+6 weeks' gestation at time of first blood draw. Nulliparous and multiparous women were both included. Participants were excluded if there was any history of nonsteroidal anti-inflammatory drug or aspirin use within the preceding 7 days, known thrombocytopenia, spherocytosis or coagulopathy. If a participant developed a pregnancy complication they were then excluded from the study. In this study, 14 pregnant participants were excluded from the pregnant cohort for the following reasons: four excluded due to difficult blood draw and small blood sample so no longitudinal data, one intrauterine fetal death, three had a preterm delivery, one developed pre-eclampsia, one developed gestational diabetes and four withdrew consent before completion of the study.

Pregnant participants were also asked to attend for repeat phlebotomy at their second-trimester ultrasound visit (between 20 and 22 weeks of gestation) and again at their 28-week or 34-week antenatal visit.

Novel platelet assay and phlebotomy

A protocol for blood draw to maintain platelet integrity was used for all phlebotomy procedures and phlebotomy was performed by a physician. Venous blood was drawn from the antecubital fossa from an uncuffed arm via a 19-gauge butterfly needle and the first 5 ml drawn was sent for complete blood count and confirmed a normal platelet count in each participant. The following 27 ml of blood was drawn into a syringe that contained 3 ml of 3.2% sodium citrate. Blood was then centrifuged for 10 minutes at 150 g. Platelet-rich plasma was aspirated and used to perform the assay.

A modification of light transmission aggregometry was used to assess platelet reactivity. This novel assay allowed five agonists to be tested simultaneously at multiple submaximal concentrations, so recreating physiological conditions. Agonists included arachidonic acid (AA), collagen, epinephrine, adenosine diphosphate (ADP) and thrombin receptor activating protein (TRAP). This assay has been described in detail elsewhere.[4-6]

Statistical analysis

Nonlinear regression was used to model platelet aggregation across various agonist doses, for each agonist. The best fit model for the dose–response curves and the half maximal effective concentration (EC50) were compared between each trimester and each trimester and the nonpregnant controls using an analysis of variance-like method (F-tests). The curves were generated and subsequent analysis was performed using GraphPad Prism Version 5 (GraphPad, San Diego, CA, USA).

Results

A total of 30 participants were included in the nonpregnant cohort. All of the participants included in this study had confirmed previous normal pregnancy outcomes. Of the 50 pregnant participants, a total of 36 participants had a platelet assay in all three trimesters and had a confirmed normal pregnancy outcome. The participant details of both groups are outlined in Table 1. The average gestational age at delivery was 39+4 weeks (range 37+0–40+8 weeks). The mean birthweight for the cohort was 3687 g (SD ± 436 g). The mode of delivery of this cohort included 66% (n = 24) spontaneous vaginal deliveries, 17% (n = 6) caesarean sections and 17% (n = 6) operative vaginal deliveries.

Table 1. Demographic details
DemographicsNonpregnant (n = 30)Pregnant (n = 36)P value
  1. a

    Maternal age differed between the groups, however, in a covariate-adjusted dose–response analysis; maternal age had no differential effect on platelet aggregation between the controls and any trimester in the pregnant group in any assay.

Age (years); mean (range)36 (28–49)32 (18–39)0.0007a
Weight (kg); mean ± SD67.2 ± 866.9 ± 100.8844
Multiparous100% (= 30)67% (n = 24)n/a
PhlebotomyProgesterone level, 4.2 ± 1.6 nmol/l

First trimester: 11–14 weeks

Second trimester: 19–24 weeks

Third trimester: 28–34 weeks

n/a
Platelets272 × 109/l

First trimester: 263 × 109/l

Second trimester: 259 × 109/l

Third trimester: 260 × 109/l

ns

Platelet reactivity

Platelet reactivity for the non-pregnant control group was measured on one occasion and the dose–response curves for AA and collagen are shown in Figures 1 and 2 and the EC50 for each agonist is listed in Table 2. Platelet reactivity was then measured on three separate occasions in pregnancy and the dose–response curves for each agonist in each trimester were plotted. Dose–response curves of percentage platelet aggregation to increasing submaximal doses of each agonist were plotted for each group. The dose–response curves for each trimester for agonists AA and collagen are demonstrated in Figures 1 and 2. Comparisons between trimesters were then performed. Significant increases in platelet reactivity were seen in response to collagen and AA, as demonstrated by the dose–response curves in Figures 1 and 2 as pregnancy progressed. The difference with respect to platelet aggregation was most significantly increased in response to collagen across trimesters.

Table 2. Compares the EC50 in the nonpregnant controls to the EC50 of collagen, AA and epinephrine in each trimester
AgonistEC50 pregnant women (95% CI)EC50 nonpregnant control women (95% CI)P value
  1. a

    A significantly different EC50 compared with the nonpregnant controls.

Collagen (μg/ml)
First trimester0.09 (0.08–0.11)0.04 (0.03–0.05)<0.001a
Second trimester0.10 (0.04–0.27)0.01a
Third trimester0.05 (0.03–0.1)0.320
Arachidonic Acid (μg/ml)
First trimester0.28 (0.1–0.75)0.35 (0.14–0.88)0.75
Second trimester0.10 (0.07–0.15)0.04a
Third trimester0.08 (0.06–0.13)0.03a
Epinephrine (μmol/l)
First trimester0.01 (0.003–0.02)0.11 (0.07–0.18)0.02a
Second trimester0.01 (0.002–0.06)0.30
Third trimester0.04 (0.02–0.14)0.05
Figure 1.

Dose–response curves for collagen plotting platelet reactivity to increasing doses of collagen. The dose–response curves compare the ‘best fit of the curve’ in each trimester with the nonpregnant controls. This graph demonstrates that platelet reactivity is significantly reduced in the first trimester compared with the nonpregnant controls (< 0.001).

Figure 2.

Dose–response curves for AA that demonstrate maximal platelet reactivity to increasing doses of AA. The dose–response curves for each trimester were compared using the ‘best fit’ analysis and plotted against –response curves for the nonpregnant control group. This graph demonstrates significantly increased platelet reactivity to AA in the second and third trimesters compared with the first trimester and the nonpregnant state (< 0.001).

Then we compared the pregnant group with the nonpregnant group. There is a marked reduction in platelet reactivity in the first trimester as represented by the dose–response curves in Figure 1 compared with the nonpregnant controls (< 0.0001). Platelet reactivity to collagen was then demonstrated to increase significantly in the second and third trimesters towards a nonpregnant baseline. The EC50 for collagen was significantly different in the first and second trimesters compared with the nonpregnant group as shown in Table 2.

The dose–response curves for AA in each trimester and the nonpregnant group are demonstrated in Figure 2. There was no difference between the pregnant and nonpregnant cohorts in the first trimester, but there was a significant difference in the second and third trimesters. There was a trend of increasing platelet reactivity with advancing gestation when comparing the second and third trimesters with the first (P < 0.0001). The EC50 also increased in the second and third trimester as outlined in Table 2.

Dose–response curves and the EC50 for ADP and TRAP did not show any significant difference between groups and no consistent trends across pregnancy. Epinephrine generated similar dose–response curves to AA but this did not reach statistical significance.

Discussion

Main findings

We carried out a comprehensive assessment of platelet reactivity in pregnancy evaluating five agonists at multiple concentrations at defined time points in pregnancy. The results of this study demonstrate that platelet reactivity fluctuates at different gestations of pregnancy, according to agonist and is significantly different compared with the nonpregnant state. There are two prominent, clinically relevant findings; first, the significantly reduced platelet reactivity in response to collagen in early pregnancy and second the significantly increased platelet reactivity in the second and third trimesters in response to AA. Although there were changes in reactivity with the agonists epinephrine, TRAP and ADP at different gestations, no consistent trends or significance was demonstrated.

Strengths and weaknesses

The strengths of this study include the use of this novel platelet assay, which allowed us to assess multiple platelet receptors at various submaximal concentrations. This is more reflective of platelet response in vivo. In this study phlebotomy was carried out in a controlled manner to maintain platelet integrity and all participants were controlled to limit the impact of lifestyle factors on platelet reactivity.[2] Platelet reactivity was assessed longitudinally in the pregnant participants to assess the influence of gestational age on platelet reactivity. However, this research would be even more valuable if the women in the pregnant cohort had acted as their own controls with an assay taken remote from pregnancy. It proved difficult to control these women for factors including breastfeeding, hormonal contraception and progesterone levels in the postnatal period. The nonpregnant group were much easier to control for confounding factors know to affect platelet reactivity but the group was significantly older than the pregnant cohort. Even when the age difference was controlled for by statistical methods it did not explain the differences in platelet reactivity between the two groups. To our knowledge there is no evidence that age affects platelet reactivity.

Many studies have used different methods for analysing platelet reactivity with varying numbers and concentrations of agonists to evaluate normal pregnancy.[7] It is therefore not surprising that there are many conflicting reports. For example, some studies have described hyper-reactive platelets or activated platelets. Holmes and colleagues[8] performed a longitudinal study and measured levels of soluble P-selectin. P-Selectin is a biomarker of platelet activation that is shed by de-granulated platelets. It was measured in each trimester of normal pregnancy using an immunoassay. A significant increase in soluble P-selectin was found in the second and third trimesters of pregnancy when compared with nonpregnant control participants. They did not examine aggregometry. Norris and colleagues[9] used whole blood impedance aggregometry to investigate platelet reactivity in a normal pregnant cohort. They found increased aggregatory responses to all agonists; AA, collagen, epinephrine and platelet-activating factor as pregnancy progressed. However, they did not have a nonpregnant control group with which to make a comparison. Morrison and colleagues[10] described hyper-reactive platelets in response to AA but they only used a single large concentration of AA. Hyporeactive platelets have also been reported although less commonly. Oron and colleagues[11] measured serum soluble CD40, a transmembrane protein shed from activated platelets, using an immunoassay. They showed a lower level of soluble CD40 in the normal pregnant cohort compared with the nonpregnant state. O'Brien and colleagues[12] also demonstrated hyporeactive platelets using aggregometry in response to a number of agonists but the pregnant population were only assessed at one time point in pregnancy. The conflicting reports in these studies arise frequently from the use of different study designs as a means of assessing platelet function. Our study provides a robust assessment of platelet reactivity because it examines five agonists at multiple submaximal concentrations with a well-controlled nonpregnant group and longitudinal pregnancy data.

A feature of this study is the reduced platelet reactivity to collagen in the first trimester, which then returns to a ‘baseline’ reactivity level in the third trimester. Collagen is one of the major adhesion molecules involved in clot formation. Primary haemostasis begins after endothelial disruption and exposed collagen leads to a sequence of events whereby platelets adhere to the exposed collagen and mediate clot formation with fibrinogen. Platelets have a role in placentation and there is evidence to suggest that the release of a number of growth factors and cytokines induce extravillous trophoblast migration at the feto–maternal interface.[13, 14] The process of maternal vascular remodelling taking place in early pregnancy exposes subendothelial collagen to platelets on a continual basis. This study reconciles these two processes; a reduced platelet response to collagen in the first trimester allows vascular remodelling without fibrinogen deposition and clot formation yet the platelets are free to circulate in the microvasculature releasing the necessary soluble factors for trophoblast invasion. Extravillous trophoblasts continue to remodel the spiral arteries until 20 weeks of gestation.[15] This study demonstrates in the second trimester an increase in platelet reactivity to collagen, which increases further in the third trimester. At these stages of pregnancy, primary haemostasis becomes far more important to control blood loss at the time of delivery.

Platelets showed an increased reactivity in the second and third trimesters to AA in this study. There are enormous changes in the process of thrombosis and haemostasis in pregnancy and a fine balance between the two must be achieved. The increase in many coagulation factors is well established including von Willebrand's factor, factors VII, VIII, X and XII.[16-19] Arachidonic acid works via the thromboxane A2 cyclooxygenase pathway to induce platelet aggregation. The increased aggregation to AA is an additional explanation for the pregnant ‘pro-thrombotic’ state where the balance towards thrombosis is favourable in later gestations to prevent postpartum haemorrhage. However, the increase in AA-induced aggregation demonstrated in this study would also suggest an increased risk of arterial thrombosis, which is important to consider in the clinical context of increasing maternal age and obesity.

This study outlines comprehensively the reactivity of platelets in uncomplicated pregnancy. It shows that during pregnancy platelet reactivity is altered compared with the nonpregnant state and that these changes are gestation and agonist dependent. It provides valuable detail about platelet physiology and forms the basis for studying platelets in complicated pregnancies such as those involving pre-eclampsia or intrauterine growth restriction (IUGR). No other research has provided longitudinal pregnancy data on multiple agonists at physiological doses. Future areas of research should examine the use of this novel platelet assay in complicated pregnancies such as those involving pre-eclampsia or IUGR. Further evaluation of the role of collagen-induced and AA-induced platelet reactivity in pre-eclampsia and IUGR may reveal valuable insights into the pathophysiological mechanisms of these disorders.

Conclusion

Platelet reactivity varies according to pregnancy state, gestational age and agonist. The finding that platelet reactivity is reduced in the first trimester of pregnancy may be useful for the interpretation of further studies examining the role of platelet reactivity in the first trimester of pregnancies that develop uteroplacental disease. The first trimester may be the most clinically valuable time to investigate agonist-induced platelet reactivity, to determine if there is any role for the assessment of platelet reactivity in the prediction of pre-eclampsia or IUGR.

Acknowledgements

We would like to acknowledge Eimear Dunne for her assistance with the laboratory work in this study.

Disclosure of interests

None of the authors have a conflict of interest.

Contribution to authorship

NB, KF, DK, MG and FM designed the study. NB, KF, MD, AM and LF recruited the participants. BC performed all of the assays, PD performed the statistical analysis and NB wrote the manuscript with help from MG, DK, FM.

Details of ethics approval

Ethics approval was obtained from the Rotunda Hospital Research Ethics Committee before commencing this study and the study complied with the Declaration of Helsinki. Full written consent was obtained from all participants before inclusion in this study. Ref no. S45C-212031217330; date 17 November 2007.

Funding

This study was supported with a grant from the ‘Friends of the Rotunda’ Registered Charity number CHY240 (Ireland).

Ancillary