Early-onset fetal growth restriction: comparison of two management protocols in a single tertiary center

Abstract Objective Compare the neonatal outcomes of two protocols of diagnosis and surveillance of pregnancies complicated by early-onset FGR in a tertiary hospital. Methods This is a retrospective cohort study of pregnant women diagnosed with early-onset FGR between 2017 and 2020. We compared the obstetric and perinatal outcomes between two different management protocols (before and after 2019). Results Seventy-two cases of early-onset FGR were diagnosed in the forementioned period: 45 (62.5%) were managed according to protocol 1 and 27 (37.5%) according to protocol 2. Mean gestational age at delivery was significantly different between groups: 34.9 ± 3.1 weeks (95% CI 34.0–35.9) in group 1 and 32.3 ± 4.4 weeks (95% CI 30.4–33.9) in group 2. 74.1% (20) of newborns in group 2 were admitted in de NICU, a significant difference when compared with 46.7% of group 1. There were no statistically significant differences in the remaining serious neonatal adverse outcomes. Conclusions This is the first study published comparing two different protocols of management of FGR. The implementation of the new protocol seems to have led to a decrease in the number of fetuses labeled as growth restricted and to a decrease in the gestational age of delivery of such fetuses, but without increasing the rate of serious neonatal adverse outcomes. Synopsis The implementation of the 2016 ISUOG guidelines for the diagnosis of fetal growth restriction seems to have led to a decrease in the number of fetuses labeled as growth restricted and to a decrease in the gestational age of delivery of such fetuses, but without increasing the rate of serious neonatal adverse outcomes.


Introduction
The term fetal growth restriction (FGR) is used to describe a fetus not reaching its endorsed growth potential. Our goal as clinicians is to identify which fetuses are at risk since many whose estimated fetal weight (EFW) lies below the 10th centile are in fact constitutionally small and therefore have a lower risk of adverse perinatal outcomes. In contrast, a growthrestricted fetus, even with an EFW above the 10th centile, has a higher risk of complications.
FGR poses many diagnostic and management challenges. The most important is the correct identification of the fetuses at risk of adverse outcomes. Since 2016, a consensus definition of FGR was established [1], which included not only the EFW but also Doppler parameters. However, many controversies remain regarding the best approach to diagnose and manage fetuses with growth restriction, with new and conflicting Society for Maternal-Fetal Medicine [2] and International Society of Ultrasound in Obstetrics and Gynecology [3] guidelines being published in 2021. The sole use of an EFW or abdominal circumference (AC) <10th as a criterion to diagnose FGR has a higher sensitivity but a lower specificity [4] which results in the labeling of a higher proportion of fetuses as growth restricted when they are only constitutionally small, leading to enhanced monitoring and potentially unnecessary interventions [5]. The combination of biometric parameters with Doppler indices has an overall higher specificity, positive likelihood ratio and positive predictive value to identify a small gestational age neonate [4].
In 2019, our center adopted the aforementioned 2016 consensus definition and management protocol for FGR.
The aim of this work is to compare the neonatal outcomes of two consecutive protocols of diagnosis and surveillance of pregnancies complicated by earlyonset FGR in a single Portuguese tertiary hospital.

Methods
We conducted a retrospective cohort study of pregnant women diagnosed with early-onset FGR at a single institution between 2017 and 2020. Patients were included if they had a singleton pregnancy and a gestational age at or above 24 weeks. We excluded women whose birth did not occur at our institution and fetuses with severe congenital anomalies.
In February 2019, the management protocol for fetuses diagnosed with FGR was updated in our institution. Before February 2019 (group 1), FGR was diagnosed if a fetus had an EFW [6] < 10th centile and/or an AC [6] < 10th centile for the gestational age and/or an abnormal Doppler (umbilical artery [7], middle cerebral artery [8] and/or cerebroplacental ratio [8]). The distinction between early and late-onset FGR was made at 34 weeks.
After February 2019 (group 2), the definition of FGR followed the previously described Delphy procedure for early-onset FGR: AC 6 or EFW 6 was < 3rd centile; or absent end-diastolic flow in the umbilical artery [7] (UA); or AC 6 or EFW 6 < 10th centile combined with a pulsatility index (PI) > 95th centile in either the umbilical [7] or uterine artery [9], diagnosed before 32 week's gestation. Tables 1 and 2 describe the management protocols and termination of pregnancy criteria.
We compared maternal demographic information and perinatal outcomes between the two protocols.
The primary outcome was a composite outcome of perinatal death (up to the discharge home from neonatal services) and/or one of the following morbidities: admission to the neonatal intensive care unit (NICU), bronchopulmonary dysplasia, intraventricular hemorrhage, neonatal sepsis, necrotizing enterocolitis, need for neonatal respiratory support or retinopathy.
Secondary outcomes included birth weight, gestational age at delivery, indicated preterm delivery, 5-min Apgar score <7 and cesarean section for fetal distress during labor.
Maternal, neonatal and obstetrical data were extracted from clinical records.
Baseline characteristics were compared between the two groups using the unpaired Student t test as appropriate for continuous variables and the v 2 test or Fisher exact test as appropriate for categorical variables. All tests were two-tailed with a p value of 0.05 considered significant. Statistical analysis was completed with IBM V R SPSS V R Statistics 26.
This study was approved by the ethics committee of the Lisbon Academic Medical Center (reference number 347/21).
A higher proportion of fetuses with early-onset FGR had a UA-PI >95th centile in group 1 when compared to group 2 (66.7% vs 40.7%, p ¼ .031). All other Doppler changes detected during surveillance were similar between both groups (Table 4).  Mean gestational age at delivery was significantly different between groups: 34.9 ± 3.1 weeks (95% CI 34.0-35.9) in group 1 and 32.2 ± 4.4 weeks (95% CI 30.4-33.9) in group 2. There were no differences regarding indicated preterm delivery, the type of delivery or the need for a cesarean-section for nonreassuring fetal status (Table 4).
Birth weight was significantly lower in group 2. For protocol 1 and 2, 78% (18) and 85% (23) of newborns had a birth weight <10 th centile, respectively. There were two neonatal deaths in each group and 74.1% (20) of newborns in group 2 were admitted to de NICU, a statistically significant difference when compared with 46.7% in group 1. More neonates in group 2 needed non-invasive respiratory support and developed grade I intraventricular hemorrhage and retinopathy of prematurity. There were no statistically significant differences in the remaining neonatal outcomes ( Table 5).

Discussion
In 2019, our center adopted the ISUOG recommendations for the diagnosis of FGR [1]. The sole use of an EFW or AC <10th as a criterion to diagnose FGR, as was used previously at our center, has a higher sensitivity but a lower specificity [4]. The combination of biometric parameters with Doppler indices has an overall higher specificity, positive likelihood ratio and positive predictive value to identify a small gestational age neonate [4]. In line with that, we observed a decreasing number of FGR diagnosis between 2017-2018 (45 cases) and 2019-2020 (27 cases), since we changed the diagnostic criteria to include Doppler indices and consequently increased the specificity of the diagnosis.
We describe a significant difference in mean gestational age at delivery between both protocols, with a lower gestational age in group 2 (32.3 ± 4.4 weeks) when compared with group 1 (34.9 ± 3.1 weeks). This important difference can put into question the relevance of our recently implemented protocol. However, this is probably related to the differences in the diagnostic criteria for early-onset FGR. Additionally, when Data are given as mean ± SD or n (%). BMI: body mass index. analyzed in detail, we observe a tendency toward a higher proportion of maternal-indicated preterm deliveries, although not reaching statistical significance, likely because of the small sample size. The four cases with maternal-indicated preterm delivery were associated with serious conditions unrelated to the diagnosis of FGR, specifically two cases of severe preeclampsia, one case of placental abruption and one case of HELLP syndrome. It is important to highlight that the indication for preterm delivery was not based on Doppler changes or suspected fetal hypoxia. However, there was also a tendency for a higher proportion of pregnancies interrupted for fetal reasons (although not reaching statistical significance) likely motivated by the changes in Doppler criteria for pregnancy interruption in the most recent protocol. The lower gestational age at delivery is responsible for the poorer neonatal outcomes, such as admission to the NICU, need for non-invasive respiratory support, grade I/II intraventricular hemorrhage and retinopathy. However, it is important to highlight that serious adverse outcomes, such as grade III or IV intraventricular hemorrhage or severe bronchopulmonary dysplasia, were rare events in both groups. Also, we emphasize that there were no stillbirths after the implementation of the new protocol.
Early-onset FGR is associated with an increased risk of neonatal morbidity and mortality [11][12][13][14][15][16][17]. A study [17] that evaluated the short-come outcomes of the TRUFFLE cohort reported a 29% rate of bronchopulmonary dysplasia and 42% rate of ventilated neonates, both of which are in line with our results after the implementation of our new protocol. On the other hand, the rate stillbirth (two percent), necrotizing enterocolitis (three percent) and grade III or IV intraventricular hemorrhage (two percent) were higher than our results (we had no cases of stillbirth, necrotizing enterocolitis, or severe intraventricular hemorrhage). Similarly, we report a lower rate of neonatal sepsis (32% vs 26%). A systematic review of 2019 [14], reports a broader rate of adverse outcomes such as bronchopulmonary dysplasia (4-19%), intraventricular hemorrhage (0-25%), necrotizing enterocolitis (0-22%), retinopathy of prematurity (2-29%) and sepsis (25-64%). Our outcomes seem to be similar to those reported in the literature.
To our knowledge, this is the first study published comparing two consecutive different protocols of management of fetuses with growth restriction. This was a single-center study with a team of clinicians that has remained stable throughout the last few years which allowed a uniform implementation of both management protocols. On the other hand, this is a retrospective study with a small sample, which impaires our ability to identify rarer adverse outcomes such as perinatal death.
The implementation of the new protocol seems to have led to a decrease in the number of fetuses labeled as growth restricted and to a decrease in the gestational age of delivery of such fetuses, but .082 Ã Perinatal death (up to the discharge home from neonatal services) and/or one of the following morbidities: admission to the neonatal intensive care unit (NICU), bronchopulmonary dysplasia, intraventricular hemorrhage, neonatal sepsis, necrotizing enterocolitis, need for neonatal respiratory support or retinopathy.
without increasing the rate of serious neonatal adverse outcomes.
The use of protocols allows for standardized decisions in a field where there are still many points that are not consensual. It is of paramount importance that more high-quality studies focus on the best diagnostic and management protocols in order to improve survival without serious morbidities of fetuses affected by growth restriction.

Author contribution
Joana Barros, Andr e Graça and Rui Marques Carvalho participated in the conception and design of this work, revised it critically and provided relevant intellectual content and approved the final version we present here.
Ana Dagge participated in the conception and design of this work, and was responsible for the acquisition, analysis and interpretation of data as well as for writing this manuscript.

Declaraion statement
No potential conflict of interest was reported by the author(s).

Funding
The author(s) reported there is no funding associated with the work featured in this article.