Early‐onset fetal growth restriction: A systematic review on mortality and morbidity

Abstract Introduction Severe early‐onset fetal growth restriction is an obstetric condition with significant risks of perinatal mortality, major and minor neonatal morbidity, and long‐term health sequelae. The prognosis of a fetus is influenced by the extent of prematurity and fetal weight. Clinical care is individually adjusted. In literature, survival rates vary and studies often only include live‐born neonates with missing rates of antenatal death. This systematic review aims to summarize the literature on mortality and morbidity. Material and methods A broad literature search was conducted in OVID MEDLINE from 2000 to 26 April 2019 to identify studies on fetal growth restriction and perinatal death. Studies were excluded when all included children were born before 2000 because (neonatal) health care has considerably improved since this period. Studies were included that described fetal growth restriction diagnosed before 32 weeks of gestation and antenatal mortality and neonatal mortality and/or morbidity as outcome. Quality of evidence was rated with the GRADE instrument. Results Of the 2604 publications identified, 25 studies, reporting 2895 pregnancies, were included in the systematic review. Overall risk of bias in most studies was judged as low. The quality of evidence was generally rated as very low to moderate, except for 3 large well‐designed randomized controlled trials. When combining all data on mortality, in 355 of 2895 pregnancies (12%) the fetus died antenatally, 192 died in the neonatal period (8% of live‐born neonates) and 2347 (81% of all pregnancies) children survived. Of the neonatal morbidities recorded, respiratory distress syndrome (34% of the live‐born neonates), retinopathy of prematurity (13%) and sepsis (30%) were most common. Of 476 children that underwent neurodevelopmental assessment, 58 (12% of surviving children, 9% of all pregnancies) suffered from cognitive impairment and/or cerebral palsy. Conclusions When combining the data of 25 included studies, survival in fetal growth restriction pregnancies, diagnosed before 32 weeks of gestation, was 81%. Neurodevelopmental impairment was assessed in a minority of surviving children. Individual prognostic counseling on the basis of these results is hampered by differences in patient and pregnancy characteristics within the included patient groups.


| INTRODUC TI ON
Severe early-onset fetal growth restriction (FGR) with placental insufficiency as its mechanism 1 is an obstetric condition that is mostly managed in tertiary-care hospitals. By consensus, FGR is defined as onset before 32 weeks of gestation, a fetal abdominal circumference or estimated fetal weight (EFW) below the 3rd centile or absent enddiastolic flow in the umbilical artery, or abdominal circumference or EFW below the 10th centile combined with a pulsatility index of the uterine artery above the 95th centile and/or pulsatility index of the umbilical artery above the 95th centile. 2 This patient group needs high amounts of care and has a high likelihood of iatrogenic premature delivery, both for fetal and for secondary maternal indications, such as the development of the maternal syndrome of preeclampsia. 3 As these FGR children are usually born very preterm, the condition carries significant risks of neonatal mortality, major and minor morbidity, and long-term health sequelae. 4,5 These risks are not only strongly related to gestational age, but also to the extent of growth restriction. Reported survival rates vary. 3 Counseling patients with severe early-onset FGR about perinatal prognosis is difficult because of the uncertain influence of different prognostic variables of the condition. Furthermore, the widespread variability of existing data on survival and long-term prognosis of the fetus makes decision-making in this patient group even more difficult.
Overview of total mortality is often lacking in literature on this subject. For example, many studies describe the prognosis of liveborn neonates after FGR and do not take antenatal death into account. From an obstetric perspective, long-term outcomes can only be interpreted optimally if they are presented together with the proportions of antenatal and neonatal death. 6 The aim of this systematic review is to describe the chances of overall (antenatal and neonatal) survival, and long-term morbidity and neurodevelopment based on the total number of fetuses at first FGR diagnosis to inform patients and obstetricians in their counseling and decision-making.  (Table S1). The retrieved records were imported and de-duplicated in endnote X7. The included studies were screened for additional relevant cited or citing references.

| Main outcomes measures
Six important research questions were identified: 1. What is, in severe early-onset FGR, the chance of intrauterine death?
2. What is, in live-born neonates after severe early-onset FGR, the chance of neonatal death?
3. What is, in surviving children after severe early-onset FGR, the chance of neurodevelopmental impairment (NDI) at or before 5 years of age in long-term follow up? 4. What is, in surviving children after severe early-onset FGR, the mean cognitive score at or before 5 years of age? 5. What is, in surviving children after severe early-onset FGR, the mean motor score at or before 5 years of age?
6. What is, in surviving children after severe early-onset FGR, the chance of cerebral palsy at or before 5 years of age?

| Eligibility criteria
Records covering singleton pregnancies diagnosed with FGR, as defined by trialists, diagnosed before 32 weeks of gestation, were included when the antenatal and perinatal data on mortality were Neurodevelopmental impairment was assessed in a minority of surviving children.
Individual prognostic counseling on the basis of these results is hampered by differences in patient and pregnancy characteristics within the included patient groups.

K E Y W O R D S
estimated fetal weight, fetal growth restriction, fetal mortality, infant mortality, morbidity, neurodevelopmental impairment

Key message
The data of 25 included studies, reporting 2895 pregnancies complicated by fetal growth restriction, diagnosed before 32 weeks of gestation indicate that overall survival was 81%. In 12% of the surviving children cognitive impairment and/or cerebral palsy was diagnosed.
reported. If a study included patients diagnosed with FGR before and after 32 weeks of gestation (for example between 24 and 38 weeks of gestation) the study was only included if data on the subgroup below 32 weeks of gestation was reported separately in the publication. Because of the progress of quality of obstetric and neonatal care, only patient groups (partially) born in or after the year 2000 were included. Furthermore, only records published in English and with an available full text were included.
Records were excluded if they only described neonates born after FGR, evaluating the postnatal data, without describing the antenatal and perinatal mortality.

| Data collection
Titles and abstracts of all search results were independently screened by 2 researchers (AP and IMB). Discrepancies were resolved by discussion with a third researcher (WG). The full text of potentially eligible studies was assessed. Relevant data were extracted from the full text by 2 researchers independently (AP and IMB) and compared for purpose of completeness and correctness.
The quality of the evidence was rated using the GRADE instrument. 7

| RE SULTS
The literature search identified 2602 unique records, and 2 additional records were identified through reference and citation checks. After title and abstract screening, 269 full-text records were assessed for eligibility; 25 studies comprising 2895 patients were included in the systematic review ( Figure 1).   [8][9][10][11][12] were judged as "unknown" risk of bias. This judgment was mostly based on the fact that these studies were retrospectively registered and not blinded, and that some of the baseline criteria and outcomes were not reported for pregnancies that involved neonatal death. The other RCTs and the observational studies included were generally judged as "low" risk of bias.
A subset of the studies report neonatal morbidity (see Supplementary material, Table S3). When combining the data, 34% of the live-born neonates experienced respiratory distress syndrome (2 studies, range 34%-36%), 9.1% had bronchopulmonary dysplasia (4 studies, range 4%-19%), 4.3% had intraventricular hemorrhage (10 studies, range 0%-25%), 5.6% had necrotizing enterocolitis (9 studies, range 0%-22%), 2.6% had persistent pulmonary hypertension of the newborn (2 studies, range 1.9%-9.1%), 12.5% had retinopathy of prematurity (4 studies, range 2%-29%) and 30% had sepsis (4 studies, range 25%-64%). One study used a composite outcome for severe neonatal morbidity 13 and 1 study used a composite for respiratory distress syndrome and chronic lung disease. 14 The ages at which the neurodevelopmental outcome was assessed, the types of tests used for the assessment and the definition of NDI differed between studies. Therefore, not all studies could be included in the evidence table. From the 476 children (402 from 1 larger study, the remainder from 6 small studies) who underwent neurodevelopmental assessment (Table 3)

| D ISCUSS I ON
The aim of this systematic review was to collate evidence on the perinatal mortality, morbidity and long-term (neuro-)development of pregnancies complicated by early-onset FGR. Particularly in pregnancies with fetal compromise around the limits of viability, information on fetal and neonatal prognosis could offer a guide in decision-making for parents and obstetricians.
We found that antenatal mortality was about twice as high as neonatal mortality. Only a few studies reported on the number of children diagnosed with relevant neonatal morbidity, such as respiratory distress syndrome, bronchopulmonary dysplasia, persistent pulmonary hypertension of the newborn and retinopathy of prematurity. Also, a minority of the studies reported outcomes of longterm follow up. Moreover, neurodevelopmental assessments were performed at different ages and different neurodevelopmental measures were used.
The strength of this systematic review is the broad literature search and the strict inclusion criteria. We excluded studies that included all their patients before 2000, as the level of (neonatal) health care was essentially different in that period. Many studies that reported long-term follow up did not include the antenatal and/ or neonatal mortality of the sample studied, 5,15 which could create selection bias and may lead to numbers on healthy survival of early-onset severe FGR to be too optimistic. Therefore, we also predefined to exclude studies that used live birth or survival as starting criteria, as we consider it crucial to include data on all-type mortality to allow proper conclusions about prognosis from the obstetric perspective. Severity of brain damage is not only associated with FGR, but also with perinatal/neonatal management, and survival bias was therefore taken into account. One weakness of this systematic review is the lack of consistency in the definition of FGR in the included studies. As is highlighted in One large well-designed RCT 20 provides high-quality evidence on the mortality and morbidity outcomes and neurodevelopmental outcomes at 2 years of age. 10 Limitations of this study are that it is a trial on patient management and some pregnancies were excluded because of fetal distress. However, the advantage of this RCT was the strict inclusion criteria of FGR and the relatively well-organized follow up with high attrition rate.
Currently, there are no specific evidence-based therapies for early-onset severe FGR. In the absence of therapeutic interven- excluded studies reporting on wider ranges of gestational age. This included 2 well-designed studies investigating long-term neurodevelopment. 6,21 In these studies, 10 out of 34 (29%) and 14 out of 149 (10%) children, respectively, had an abnormal IQ score, of which the latter percentage is in line with the findings of this systematic review. Together with the studies included in our analyses that reported on long-term neurodevelopment, it illustrates the need for more prospective studies starting at diagnosis of FGR and extending to early school age development of the surviving children.

| CON CLUS ION
In this systematic review based on 25 studies comprising 2895 pregnancies complicated by severe early-onset FGR, we found that the overall rates of antenatal and neonatal death were 12.3% and 6.6%, respectively. Of the 476 children included in the long-term follow up, 12.2% of the survivors (7.9% of all pregnancies) were affected by NDI and/or cerebral palsy. Data on neurodevelopment were much less reported and mostly during toddler years, and not school age.
Conclusions at an individual level are hampered by the differences in study quality and prognostic characteristics. A future analysis with individual patient data might further improve individual patient counseling. Longer follow up in prospective FGR cohorts is needed to provide data on the balance between mortality and NDI.