The impact of intrauterine growth restriction (IUGR) on neonatal primary hemostasis

Platelet function in IUGR neonates remain a eld of debate, especially in preterm subgroup. Platelet Function Analyzer (PFA-100) offers a quantitative in vitro assessment of primary, platelet-related hemostasis. Our aim was to examine platelet function using PFA-100 in term and preterm IUGR neonates and associate our results with several perinatal parameters. PFA-100 was applied on 74 IUGR neonates, 48 full-term (> 37 weeks’ gestation) and 26 preterm neonates (< 37 weeks’ gestation). Control group consisted of 118 healthy neonates. Two CTs (with COL/EPI and COL/ADP cartridges) were determined on cord blood samples for each subject. COL/EPI CTs were prolonged in IUGR (median 132s) compared to control neonates (median 112,5s), p=0.0371. Median COL/EPI CT for term and preterm IUGR neonates was 126s and 137s respectively (p=0.0013), and COL/ADP CT was 70s for term and 75s for preterm IUGR neonates (p=0.0827). COL/ADP CTs were shorter in IUGR neonates born via vaginal delivery (p=0.0065), shorter after intrapartum antibiotic prophylaxis with ampicillin (p=0.0045) and prolonged in neonates whose mothers received epidural anesthesia (p=0.032). The cause of IUGR has no impact on CTs (p>0.05 in all cases). COL/EPIand COL/ADP CTs had positive correlation (r=0.37, p<0.0012) whereas no correlation was proved between both CTs and hematological parameters in IUGR neonates. Conclusions: IUGR neonates showed impaired platelet function, with preterm IUGR neonates confronting the greater risk for bleeding tendency. Prolonged COL/EPI CTs seemed to be independent of hematological parameters, especially thrombocytopenia of IUGR neonates. studies among neonates mainly neonates born to mothers with preeclampsia. Several studies report decreased platelet adhesion and expression of glycoproteins on activated platelets’ surfaces. Studies using ow cytometry report higher in vitro platelet responsiveness.


Introduction
Intrauterine growth restriction (IUGR) refers to the failure of fetus to reach its intrinsic growth potential, due to pathologic causes of maternal, fetal, placental, or genetic origin [1][2][3]. IUGR affects 5-10% of all pregnancies per year [4] and is one of the leading causes of perinatal and neonatal morbidity and mortality associated with long-term chronic diseases [3,5,6]. A main characteristic of IUGR neonates is thrombocytopenia with an incidence 31.5%, three times higher compared to appropriate for gestational age (AGA) neonates [7][8][9][10][11][12]. Thrombocytopenia could be attributed to platelets' destruction caused by placental vascular pathology [13], shunt of stem cells to erythropoiesis due to intrauterine hypoxia [14], or immaturity of liver and spleen due to redistribution of blood ow and the brain-sparing effect. Early-onset thrombocytopenia is independently associated with lower gestational age (GA) at birth in IUGR neonates [8].
The investigation of neonatal platelet function in IUGR neonates remains an issue of ongoing research with con icting results so far. Platelet function studies concern mainly neonates born to mothers with pregnancy induced hypertension (PIH), showing decreased platelet adhesion [15] and expression of glycoproteins on activated platelets' surfaces [16]. On the contrary, ow cytometry studies reported higher in vitro responsiveness of neonatal platelets to various agonists [17], while studies using thromboelastometry (TEM) did not report any difference in maximal clot rmness (MCF), a measure of platelet number and function [18]. Regarding preterm IUGR neonates, data examining platelet reactivity are scarce, with platelets from preterm offsprings of PIH-pregnancies displaying lower platelet adhesion on cone and platelet analyzer (CPA) [15].
In this study we hypothesized that IUGR neonates have a distinct platelet function, possibly affecting bleeding parameters and subsequent thrombotic risk. Thus, we aimed to examine platelet responsiveness using PFA-100 in this group of neonates and associate our results with several perinatal parameters.

Materials And Methods
This is a prospective cohort study of full-term and preterm IUGR neonates born at Aretaieio Hospital, National and Kapodistrian University of Athens, during a 2-year study period (January 2017-December 2018). The Hospital Ethics Review Committee approved the study and mothers signed an informed consent prior to recruitment. Seventy-four (74) IUGR neonates, 48 full-term (> 37 weeks' gestation) and 26 preterm neonates (< 37 weeks' gestation) were included. Control group consisted of 118 AGA neonates that were previously published by our research team [19].
Demographic data and perinatal parameters were listed from the maternal and neonatal medical records and are summarized in Table 1. Neonates were monitored till discharge and any postnatal complication was recorded. IUGR group included neonates with a prenatal diagnosis of estimated fetal weight (EFW) <10th centile and a distinct IUGR causative pattern. The AUDIPOG computer-generated program was used to calculate the customized percentile for each pregnancy. Normal birth weight (BW) percentile limits were adjusted using signi cant determinants of BW (maternal age, height and weight, GA, parity and gender). (https://www.audipog.net/Estimation-croissance) Intrauterine follow up of IUGR fetuses included repeated Doppler studies assessing the pulsatility index (PI) of the uterine, umbilical, and cerebral arteries. In 22 cases PI values were in the upper limits for the corresponding GA indicating compromised fetal perfusion [20,21]. Additionally, the cephalization index (CI) (ratio of head circumference to body weight) [22] and HC/AC ratio were also used as measures of fetal compromise [23]. Exclusion criteria included cord blood Hct< 35%, cord blood platelet count < 100.000/mL, hypothermia <35 o C or any major chromosomal anomaly. higher Hct (p<0.001). Median platelet count in IUGR newborns was 243 x 10 9 /L, lower than the average platelet count in the control group 251 x 10 9 /L, although not statistically signi cant.
Additionally, differences between term (N=48) and preterm (N=24) neonates among the group of IUGR neonates were tested. According to our results COL/EPI CT ranges for term and preterm IUGR neonates were 90-157s and 104-203s respectively, and accordingly COL/ADP ranges were 62-80s (term) and 68-82s (preterm) ( Table 1). Difference was found only in COL/EPI CT (p=0.001). (Figure 2) Moreover, preterm IUGR neonates had lower median BW(p<0.001), BW centile (p<0.001), Apgar score at 1st minute (p=0.02), median HC (p<0.001), median AC (p<0.001) and higher CI (p<0.001) compared to term IUGR neonates. A higher percentage of in vitro fertilization (IVF) (p=0.03), caesarian section (CS) (p=0.006), aspirin (p<0.001) and levothyroxine administration (p=0.01) were observed among mothers of preterm IUGR neonates. (Table 1) A mixed model to control the effects of maternal aspirin administration in neonatal COL/EPI and COL/ADP CTs was implemented. Results showed that aspirin administration had no effect in COL/EPI or COL/ADP CTs neither in the complete population, nor in the IUGR or control group, nor when adjusting for term and preterm neonates (p>0.05 in all cases  Discussion IUGR neonates confront great risk of perinatal morbidities and thus platelet related primary hemostasis needs to be carefully evaluated. To the best of our knowledge, our cohort represents a study of platelet function by PFA -100 in the largest number of well-de ned IUGR neonates (n=74) reported so far. In order to make reasonable interpretation, all data and laboratory ndings were compared to a control group of 118 healthy neonates. Additionally, the comparison of special aspects between preterm and term IUGR neonates highlights important characteristics of prematurity, as preterm neonates encounter for 35% of IUGR neonates. The impact of several parameters during pregnancy and delivery of an IUGR fetus on platelet function was also tested.
According to our study lower BW and BW centile de ne IUGR neonate. The higher rate of preterm births in IUGR neonates is reasonable, as the risk of prematurity in IUGR is 3-fold greater than in AGA infants [28]. The well-known risk of IUGR infants for irregular thermoregulation [29] explains the lower temperature observed in IUGR group. IUGR infants have greater risk of perinatal stress due to a sentinel event superimposed on chronic fetal hypoxia from placental insu ciency [3] and this explains lower Apgar scores in IUGR newborns, as in our study.
As expected, preterm neonates had lower somatometric parameters (BW, BW centile, AC and HC). Furthermore, higher CI and lower Apgar scores re ect the greater compromise of prematurity. A higher percentage of preterm births was observed in pregnancies conceived by IVF, in accordance with current literature [30]. Although there is no consensus in the optimal delivery method for a preterm birth, even in IUGR fetuses, our institution's practice favors CS as depicted with the higher percentage of preterm neonates born via CS [31,32]. The higher incidence of preterm labor in pregnant women with hypothyroidism despite levothyroxine replacement therapy [33], supports higher percentage of preterm birth in our group of IUGR neonates.
Regarding platelet function in our cohort study, COL/EPI CTs were prolonged in IUGR compared to control neonates, whereas no corresponding difference was found for COL/ADP CTs. According to the helpful algorithm proposed by Favaloro [34], prolonged COL/EPI CTs with normal COL/ADP CTs could be attributed to drug effect, low Hct, mild thrombocytopenia and mild platelet/von Willebrand factor (VWF) dysfunction. Αspirin is well known to prolong COL/EPI CT, when COL/ADP CT is usually normal [26]. It is worth noticing that 20% of mothers of our IUGR neonates received aspirin during pregnancy, in contrast to 8% of mothers of control newborns. This nding complies with the recommendation of low-dose aspirin prophylaxis, initiated before 16 weeks gestation for women at high risk of preeclampsia [35]. Subanalysis and linear regression model analysis support that aspirin administration during pregnancy had no effect on COL/EPI or COL/ADP CTs, neither in the complete population nor in the IUGR or control group. A small amount of acetylsalicylic acid reaches fetal circulation [36] and results in reduced levels of thromboxane B 2 in umbilical cord [37,38]but, according to our ndings and previous platelet aggregation studies, neonatal platelet function does not appear to be affected [39,40].
It has been shown that lower Hct is correlated with prolongation of COL/EPI CTs [34], while higher neonatal Hct seems to explain shorter bleeding time (BT) and CTs compared to adults, partly by rheologic effects [41]. IUGR neonates in our study were characterized by higher Hb and Hct as polycythemia is common in IUGR neonates (up to 50%) [42]. IUGR fetuses are subjected to relative hypoxia [43] that likely triggers red cell production [44][45][46][47]. Thus, the prolongation of COL/EPI CTs among IUGR neonates of the present study could not be attributed to Hct effect. In fact, we could assume that higher Hct of IUGR neonates could not compensate platelets' impaired activation.
Low platelet counts were correlated with prolongation of CTs and should be taken into consideration when interpret CTs [24, 48,49]. For this purpose, a cut-off value of 100 x 10 9 /L was set for our study group. It is well known that the incidence of thrombocytopenia in IUGR neonates is higher [7,8] and is more prominent as the severity of IUGR increases [7,[9][10][11]. However, the median platelet count of our IUGR group (243 x 10 9 /L), which is in accordance with previous reports [9], did not differ signi cantly from the corresponding of our control group (251 x 10 9 /L), so we could not assume that thrombocytopenia of IUGR neonates explains the observed prolongation of COL/EPI CTs.
Prolongation of COL/EPI CT of IUGR neonates seems to be the result of platelet dysfunction, although this remains a eld of further investigation for this group of neonates as the majority of studies concerning platelets' activation in IUGR neonates were conducted on the subpopulation of neonates born to mothers with PIH. A study of CPA reports lower platelet adhesion in IUGR infants [15] and another study supports that preeclampsia in uences the expression of GPs on activated neonatal platelet-surface which may affect platelet function, leading to an additional bleeding risk in thrombocytopenic neonates [16]. A study showed that platelets of neonates of preeclamptic mothers had a markedly higher responsiveness to agonists in vitro by ow cytometry, however resting platelets of IUGR neonates seemed to be in a slightly lower state of activation [50]. Prolongation of COL/EPI CT in the present study supports the impaired platelet function of IUGR neonates reported from previous studies [15, 16,50]. It is of great importance that platelet hyporesponsiveness seems to apply for the whole group of IUGR neonates irrespective of the cause.
As far as white blood cell count is concerned, neutropenia is frequent in offsprings of pregnancies complicated with PIH [46,51]. IUGR neonates of our study were also characterized by lower median values of total leucocyte count compared to control group. However, leucocytes do not have any in uence on CTs [52,53].
Regarding the role of prematurity in our IUGR group, COL/EPI CT of preterm IUGR neonates was prolonged compared to term IUGR ones. COL/ADP CT of preterm IUGR neonates was also prolonged but not signi cantly. Multivariate linear regression analysis supported that prematurity affects COL/EPI CT whereas no corresponding relation was found for COL/ADP CT. Concerning low-dose aspirin administration, despite the fact that more preterm IUGR neonates were exposed in utero to aspirin, our subanalysis supports that aspirin administration during pregnancy had no effect on COL/EPI or COL/ADP CTs as previously mentioned. Preterm and term IUGR neonates did not differ in any hematological parameter (  [59]. The available evidence suggests that the platelet hyporeactivity is less well compensated by other factors, such as high Hct, in preterm compared with full-term neonates. Thus, the present study highlights that platelet dysfunction that accompanies prematurity applies also in IUGR neonates. This was also reported in offsprings of pregnancies complicated by PIH with evidence of lower platelet adhesion on CPA, meaning impaired platelet function [15]. The observation that the difference in responsiveness between platelets of IUGR and control neonates is apparent to some agonists (epinephrine) rather than others (ADP) raises several questions concerning neonatal platelet function. An increase in the available erythrocyte ADP caused by the higher Hct in IUGR neonates may compensate for the decreased ADP secretion from neonatal platelets and explains the lack of difference in COL/ADP CTs [41]. Furthermore, a ow cytometry study showed that platelets of neonates of preeclamptic mothers had markedly higher responsiveness to ADP as agonist, after in vitro activation [50].
It is worth noticing that COL/ADP CTs were found to be shorter in IUGR neonates delivered via vaginal delivery and this could be attributed to the stimulant effect of cytokines which are expressed in higher levels in vaginal deliveries [60][61][62]. Both IL-6and IL-1β induce platelet activation and aggregation through several mechanisms [63][64][65][66][67].
As far as the role of anesthesia is concerned, COL/ADP CTs were prolonged in our IUGR neonates whose mothers received epidural anesthesia during labor. Opioids seem to have no adverse effect on platelet activation and aggregation. On the other hand, ADP-induced platelet aggregation decreases with bupivacaine [68], Platelet-Activating Factor (PAF) concentrations decrease with lignocaine [69] and platelet aggregation is inhibited by ropivacaine via PFA-100 [70].
Although previous studies support that ampicillin inhibits ADP-induced platelet aggregation in adults [71,72], VLBW neonates [73] and NICU patients [74] via several mechanisms [75], the contrary was reported in our study. A possible explanation could be the small number of neonates whose mothers received ampicillin peripartum (5 out of 74). Moreover, IUGR neonates of our study did not have any evidence of bacterial infection as opposed to neonates receiving ampicillin in previous studies [73,74].
Present study presents some limitations. There were no matching venous samples to cord blood samples as our policy was not to perform unnecessary tests. Moreover, results should be interpreted in a larger number of neonates.
In conclusion, the present study supports that IUGR neonates could be at a greater risk for bleeding tendency, as indicated by the observed impaired platelet function. This platelet hyporesponsiveness is overexpressed in preterm IUGR neonates, a group that confronts mostly the risk of intraventricular and pulmonary hemorrhage. In clinical practice, the evaluation of cord blood CTs could have a prospective character; cord blood samples offer the advantage of collecting larger blood volumes and perform several tests (ie. blood gases, blood group, PFA -100), avoiding frequent neonatal blood sampling, a