Anouk Pels

This thesis aimed to answer questions on important aspects of early-onset FGR: its definition, prognosis and management.

In Chapter 2 we describe which definitions of fetal growth restriction in the existing literature are used over time and our findings underline the need for a uniform definition. In our literature review of used definitions in the years 1994, 2004 and 2014(1) we showed that through the years an increasing number of studies have been published on FGR (56, 75 and 116 respectively). Many definitions of FGR are used, of which the majority in all time frames was ‘birthweight below the 10th percentile’, which is unfitting from an obstetric perspective. This has shown improvement over the years: in 2014 the highest percentage of the studies used antepartum findings instead of postpartum findings (neonatal weight): 34% in 1994, 30% of the studies in 2004 and 47% in 2014 used antepartum findings in defining FGR. Still, the majority of definitions essentially defined SGA, i.e. smallness, rather than the pathological syndrome of FGR. This chapter points out the lack of heterogeneity and imperfections in the definition of FGR and we conclude that in order to ensure adequate interpretation from a clinical perspective as well as data synthesis from a research perspective, a uniform definition of FGR is necessary.

Chapter 3 is a systematic review on the reported fetal and neonatal mortality and short- and long-term morbidity in cohorts of women with early FGR(2). 21 Studies reporting 2334 pregnancies with FGR, diagnosed before 32 weeks of gestation, on fetal and neonatal morbidity in combination with short or long-term outcome were included in the review. Patients included in the different cohorts, had variable pregnancy characteristics such as gestational age and estimated fetal weight at diagnosis of FGR. Overall, 12% antenatal deaths and 7% neonatal deaths occurred. Only a few studies focused on the long-term neurodevelopment of the surviving children, with an overall neurodevelopmental impairment rate of 11%, but the variation was large, as were the applied methods. We conclude that, although it is obvious that early-onset FGR carries a high burden of disease for affected children, to improve counseling of individual patients about the fetal prognosis, a more detailed analysis of individual patient data would be useful.

The neurodevelopmental outcomes at five years of age in a cohort of children born after severe early-onset FGR is reported in Chapter 4(3). These children were born from mothers participating in the Trial of Randomized Umbilical and Fetal Flow in Europe (TRUFFLE) study(4). Of the 74 children that were assessed at five years of age, the mean full-scale IQ (FSIQ) was normal, but 15% of the children had an abnormal score (FSIQ below 85) and in 35% either verbal or performance or processing speed quotient was below a score of 85. Of the assessed children, 38% had an abnormal score in the motor assessment. The factors associated with an abnormal FSIQ were the end-diastolic flow (EDF) of umbilical artery, gestational age at delivery, birthweight and neonatal morbidity. The abnormal motor score was found more often in boys and in children that experienced bronchopulmonary dysplasia (BPD) in the neonatal period. In summary, the findings at five years showed an overall better outcome than anticipated, but higher risk of an adverse neurodevelopmental outcome remained, in particular the motor and processing speed outcomes. Because of the selective sample no conclusion could be made regarding the initial aims of the TRUFFLE trial.

Chapter 5 is a systematic review of the prognostic accuracy of short-term variation (STV) of the fetal heart rate on the cardiotocography (CTG) in pregnancies complicated by FGR(5). This review shows that in pregnant women with early-onset FGR (before 32 weeks of gestation), the STV is not statistically significant associated with acidemia. Due to the limited data, no conclusion can be drawn for the association between STV and fetal/neonatal mortality, morbidity and long-term (neurodevelopmental) outcomes. Therefore, we conclude that even though the STV seems a logical and promising tool for deciding on the optimal moment of delivery during fetal surveillance, a randomized controlled trial (RCT) is necessary to investigate whether a management strategy that includes a decision algorithm based on STV improves outcomes over visual evaluation of the CTG.

In Chapter 6 a secondary analysis of the Control of Hypertension In Pregnancy Study (CHIPS)(6) is presented(7). In this analysis the association has been investigated between level and duration of blood pressure control and fetal growth. The associations between gestational age at randomization to ‘less tight’ control versus ‘tight’ control of hypertension in pregnancy and the proportion of children with a birthweight below the tenth percentile were explored. The results of this analysis show that less tight control starting before 24 weeks was associated with fewer babies with a birthweight below 10th centile but a higher rate of preterm birth. There was no effect on perinatal death or high-level neonatal care for more than 48 hours. For the mother, ‘less tight’ (vs. ‘tight’) control was associated with more severe hypertension at all gestational ages, but in particular before 28 weeks. This leads to the hypothesis that ‘tight’ control of hypertension might cause some reduction in birthweight. However, this association was not powered for in the RCT and further research would be necessary to explore this hypothesis.

Chapter 7, 8, 9 and 10 are on the STRIDER (Sildenafil TheRapy In Dismal prognosis Early-onset fetal growth Restriction) trials, in particular the Dutch STRIDER study. This is a set of five trials that were designed in international collaboration, as described in Chapter 7. The aim was to investigate whether the phosphodiesterase 5-inhibitor sildenafil would decrease fetal and neonatal morbidity, mortality and long-term neurodevelopmental impairment in pregnant women with severe early-onset FGR, compared to placebo. The international study protocol of the five STRIDER trials is presented in chapter 7(8).

Chapter 8 consists of the detailed statistical analysis plan for the Dutch STRIDER trial(9).

Chapter 9 reports the results of the Dutch STRIDER trial(10). The trial was stopped early based on advice of the Data Safety Monitoring Board (DSMB) due to serious concern of harm combined with futility at interim analysis and at that time 216 patients had been randomized. The primary outcome (mortality or survival with serious neonatal morbidity) occurred in the offspring of 65 participants (60%) allocated to sildenafil and in 58 (54%) allocated to placebo (RR 1.11, 95% CI 0.88 to 1.40; P=0.38). The conclusion of this trial was that antenatal sildenafil administration, compared to placebo, did not reduce the chance of neonatal mortality and morbidity. Moreover, an unexpected and yet unexplained increase in neonatal pulmonary hypertension was observed in the infants born in the sildenafil group compared with placebo (16 (19%) compared with 4 (5%) (RR 3.67, 95% CI 1.28 to 10.51; P=0.008)).

Chapter 10 describes the process and outcomes of neonatal outcome validation within the Dutch STRIDER trial and discusses the hypotheses on the possible association between antenatal sildenafil treatment and Persistent Pulmonary Hypertension of the Neonate (PPHN). This data validation found a total rate of 12% pulmonary hypertension in the Dutch STRIDER trial. The lack of a standardized definition of pulmonary hypertension makes the diagnosis and interpreting data complex. There is need for a consensus definition of pulmonary hypertension to reduce the reported variance of pulmonary hypertension incidences in prospective studies in which neonates born preterm after FGR are actively screened for right to left shunting after 24 hours of life in order to be able to start treatment in an early stage of disease.