Current consensus on the management of fetal growth restriction

Fetal growth restriction (FGR) refers to a fetus that has failed to reach its biological growth potential. Clinically, FGR is described as early or late-onset. Detecting signs of growth deterioration and determining the appropriate timing of fetal rescue delivery is challenging. In early-onset FGR, problems are related to prematurity whereas in late-onset FGR difficulty in diagnosis and resulting fetal mortality are the key problems for the clinician. This review summarizes pathophysiology, recent consensus on the classification of clinical phenotypes, screening, diagnosis, assessment of deterioration and surveillance, monitoring intervals, the timing of birth, neurodevelopment outcomes, term fetal growth restriction and place of cerebroplacental ratio in fetal growth restriction.


Introduction Optimal growth and suboptimal growth
Optimal fetal growth, i.e. a growth that is not too small and not too big, is classified as an estimated fetal weight (EFW) between the 10th and 90 th centiles based on population-specific birth weight centiles corrected for gestational age at delivery, parity, and fetal sex for ease of detection and management. With this definition, growth restriction is an EFW below the 10 th centile, better described as small for gestational age (SGA). Fetal growth restriction (FGR) has also been described as the process where a fetus that has a certain growth potential based on genetic criteria does not attain that growth because of environmental influence ['pathological smallness'] 1 . It is distinct from the term SGA Commentary which describes both 'pathological and constitutional smallness' 1 . Since fetal weight depends on its gestational age, fetal dating scan has to be correct to prevent discrepancies from inaccurate dating.
SGA fetus carries an increased risk for stillbirth, and antenatal detection and recognition is inaccurate and suboptimal in most 2 . Antenatal recognition of FGR can halve stillbirth risk through earlier delivery of affected fetuses [fetal rescue delivery] 2 .
This review addresses pathophysiology, clinical phenotypes, surveillance and current consensus and new insights into the management of FGR.

Early-onset FGR
The classical concept is poor placentation leading to insufficient delivery of nutrients and oxygen to the fetus 1 . Uterine spiral arteries are physiologically remodelled into dilated inelastic tubes escaping maternal vasomotor control between 8-18 weeks during placenta formation, which results in a low-resistance uteroplacental circulation.
The pathophysiological pathways for defective placentation include immunological factors, endogenous vascular factors, and thrombogenic factors amongst others 1 . Poor placentation is associated with early-onset phenotype of both preeclampsia and FGR. Most of the women presenting with maternal hypertensive disorders before 34 weeks gestation will also have concurrent FGR. Sharing of a common placental pathology in both conditions may explain this phenomenon and thus confirming the concept of poor placentation 1 .

Late-onset FGR
In growth restriction presenting at later gestations, the previous association becomes less obvious. At or near term, most neonates from mothers with preeclampsia are of normal birth weight. Thus, both late FGR and late preeclampsia may be due to a different pathology 1 .
Because the term fetuses have less placental reserve capacity and require comparatively more oxygen and nutrients, the interval between the onset of the disease and the subsequent adverse outcome is shorter and less predictable 1 .
Late-onset FGR is a challenge to predict and harder to distinguish from pregnancies with normal growth. The defective placental maturational process leading to placental hypo-perfusion and late-onset FGR is difficult to diagnose because the size of the fetus and accompanying parameters may not be abnormal and further complicated by combinations and intermediate types 1 .

Clinical phenotypes
FGR evolves from a preclinical phase to a clinically obvious phase of growth delay and may eventually lead to fetal hypoxia 3,4 . Reduced growth due to decreased transplacental nutrient supply, gas transfer, and waste removal affect reserves, reflected in reduced fetal liver size and consequently abdominal circumference (AC) first, and subsequent head growth and finally affecting the entire body 3,4 . Abnormal placental perfusion in the maternal compartment results in increased blood flow resistance in the uterine artery, which is reflected in the Doppler flow-velocity waveform 3,4 . Abnormal perfusion of the fetal villous vascular tree is associated with decreased umbilical artery end-diastolic velocity (EDV) proportional to the degree of flow impairment 3,4 . Abnormal oxygen diffusion across the villous membrane leading to lower fetal arterial oxygen saturation (PaO 2 ) is associated with an attempted increase of blood supply through a decrease in middle cerebral artery (MCA) blood flow resistance 3,4 .
Management challenges in early FGR and coexisting maternal hypertensive disease include iatrogenic prematurity whereas, in late FGR, difficulties in diagnosis and surveillance leading to unanticipated stillbirth are the primary issues.

Screening
This is based on clinical findings correlated with ultrasound fetal biometry. Traditional clinical assessment of risk is based on medical and obstetrics history, abdominal palpation for fetal size and symphysis-fundal height (SFH) marked serially over a centile chart. Risk assessment can be done through detailed history according to a recommended standard such as RCOG algorithm or Growth Assessment Protocol algorithm 2,5 . Leopold's Maneuvers are used to estimate the fetal size. However, its ability to predict fetal weight is low, only 30% SGA fetuses will be detected 6 .
SFH is the most commonly used screening tool for SGA. SFH should be plotted against the gestational age starting from 24 weeks, then 28 weeks, then two weekly up to 36 weeks and weekly thereafter. SFH has a low sensitivity of 27-76% in predicting SGA, which means it potentially fails to identify over 70% of pregnancies affected by SGA 7 . This is important to consider in counselling of pregnant women 7 . In clinical practice, the SFH measurement should not be carried out in isolation and the combination of other clinical findings, medical conditions, and previous obstetric history, together will contribute to estimating the likelihood of being at risk for SGA. A higher specificity (>80%) of SFH indicates that few pregnancies which are not characterized by SGA, are referred for ultrasound evaluation 7 .
Serial SFH helps to predict the need for ultrasound parameters by lagging of fundus in a plotted chart with gestation specific centile curves. The sensitivity of SFH can be improved to 48% by using the customized charts [adjusted for physiological variables such as height, weight, parity, and ethnic origin] for the particular population, which reduces referrals for ultrasound assessment 2 . Customized charts also improve EFW assessment and reduce the false-positive diagnosis of SGA 2 .
World Health Organization (WHO) charts, GROW (Gestation Related Optimal Weight) charts, INTERGROWTH-21 st project is all different project towards standardizing of customized charts. WHO and GROW charts are widely accepted in clinical practice 2 . GROW charts are useful when looking at subgroups within the population, e.g. maternal size. There is a better correlation between SGA defined by customized centiles and perinatal mortality than any population-based standard. GROW charts also improve the definition of SGA in multi-ethnic populations by reclassification of fetuses as normal growth according to customization 8 . These reclassified fetuses had the same perinatal mortality risk as to the general, non-SGA population further emphasizing the value of customized charts 8 .
The INTERGROWTH-21 st project has been introduced as a one-size-fits-all standard for birth weight 9 . The INTERGROWTH-21 st project was a large-scale, multicountry study that measured fetal growth of babies in utero and at birth, in mothers in whom environmental, social, medical and pregnancy conditions were optimal 10 . It followed low risk, well-nourished mothers, in eight countries and was therefore meant to provide an optimal standard for all pregnancies 9,10 . The INTERGROWTH-21 st findings challenge the acceptance of customized fetal growth charts in the UK and elsewhere 10 . Adjusting growth expectations for ethnicity has no biological basis and risks normalizing suboptimal growth in disadvantaged groups 10 . However, this concept has been challenged on theoretical grounds 1,11 and when applied to detailed multi-ethnic maternity data sets, it performed poorly compared to a customized standard 2,12 .

Diagnosis
Serial ultrasound biometry and Doppler to confirm compromised circulation is the gold standard to diagnose FGR. Once diagnosed, monitoring is from umbilical Doppler flow (UAD), ductus venosus (DV) and middle cerebral artery (MCA) blood flow patterns. The combination of fetal biometry and UAD is the best available tool for the identification of an FGR fetus 5 . Uterine artery Doppler (UtAD) screening can be performed in 'high risk' cases [Thrombophilia, Previous FGR/Pre-eclampsia]. UtAD offers early evidence to prescribe low-dose aspirin in the first trimester and 85% sensitivity to predict SGA in the second trimester (22-23 weeks) in high-risk women. However, in 'low risk' mothers and after 24 weeks, it has limited predictive accuracy. Diagnosis can be based on fetal biometry alone or by considering UAD or MCA Doppler indices. Abdominal circumference (AC) less than 10 th centile has the highest sensitivity for the diagnosis of SGA, whereas EFW less than the 10 th centile is more specific 13 . Therefore serial EFW measurements are better than serial AC in detecting Commentary additional FGR, 13 . However, this has to be balanced against the potential +/-10% margin of error in estimating EFW at term using ultrasound 2 .
The diagnostic cut-off for SGA is generally accepted as below 10 th centile. One disadvantage of this cut-off is the inclusion of a variable number of constitutionally small ("normal") fetuses that do not require surveillance 2 . These otherwise normal constitutionally small babies have comparable perinatal mortality to that of the normal population 2 . EFW less than 3 rd centile or a decreased AC is more likely to identify "true FGR". The disadvantage of taking 3 rd centile for cut-off is less severe form of FGR at risk for deterioration will be missed with the resulting risk of increased stillbirth. A combined consensus has been agreed on for diagnosis 14 .
Although placental pathology does not support such clear-cut division, expert opinion considers FGR presenting before 32 weeks as "early-onset" and after 32 weeks as "late-onset" (Table 1). An estimated fetal weight less than the tenth percentile in association with an elevated umbilical artery Doppler index, a decreased middle cerebral artery Doppler index or a decreased cerebroplacental ratio (CPR) should be considered as evidence of FGR 3 . Early-onset and late-onset FGR represent two distinct clinical phenotypes of placental dysfunction. According to the Internal Society for Ultrasound in Obstetrics and Gynaecology (ISUOG) definition of FGR, "AC/EFW crossing centiles" can be defined as crossing> 2 quartiles or 50 th centile as long as it occurs in a SGA fetus (AC/EFW < 10 th centile) and/or in combination with a low CPR or elevated UAD Pulsatility Index 14 .

Assessment of the degree of fetal deterioration
Gestational age is a critical determinant in decision making. Fetal surveillance tests are applied in pregnancies with suspected FGR to estimate the risk for hypoxia, prelabor acidosis, and antepartum stillbirth, as well as the clinical deterioration. An accurate estimation of hypoxia/ acidosis is important to predict fetal compromise, risk of stillbirth and therefore the timing of delivery 3,4 . Doppler indices and biophysical profile (BPP) provide estimates for pre-labor acidosis.
Integrated multimodal fetal testing is probably the way forward compared to the elusive single best test of fetal well-being. Tests include kick count charts, clinical examination, Doppler parameters, and BPP. Abnormal umbilical flow patterns indicate an increased risk of hypoxia and acidosis. Fetal acidosis is directly proportional to the reduction in umbilical artery EDV with acidosis highest in reversed flow in end-diastole. However umbilical artery Doppler alone as a single test is not justified due to the wide range of pH distribution and additional Doppler indices are indicated depending on the gestation.
DV Doppler and presence of 'a' wave for forwarding blood floware indicated up to 32 weeks for fetuses with absent or reversed umbilical artery Doppler indices. MCA Doppler has limited accuracy to predict acidosis before 32 weeks and adverse outcomes and should not be used to time delivery in early FGR. Beyond 32-34 weeks, the umbilical artery waveform may be normal with developing acidosis. Therefore, better predictors for fetal hypoxia and deterioration is MCA PI and cerebroplacental ratio (CPR) beyond 32 weeks, which is considered as abnormal if less than 5 th centile.
BPP shows a reliable and reproducible relationship with the fetal pH, irrespective of gestational age 3,4

Selection of the monitoring intervals
Fetal surveillance is indicated to prevent stillbirth and irreversible fetal deterioration by early detection of prelabor acidosis and carry out a fetal rescue delivery balancing against neonatal morbidity and mortality of preterm birth. The consensus is based on expert opinion due to a lack of high-quality evidence to guide practice 3 . Once FGR is diagnosed, longitudinal surveillance starting at 24 to 26 weeks with integrated multimodal fetal testing, including multi vessel Doppler assessment, fetal heart rate analysis, and BPP is considered best practice 3,4 . It is expected that the combination of these tests improves the prediction of acidosis and stillbirth compared with single tests as outlined above.
In early FGR, fetal needs are less and allow a relatively chronic course leading to a small baby. Fetal deterioration typically evolves from abnormal umbilical artery Doppler studies to brain-sparing, then abnormal ductus venous Doppler, abnormal cCTG, and finally, abnormal BPP. The rate of progression is determined by the interval between diagnosis to loss of umbilical artery EDV which typically takes four to six weeks. Once forward velocities in the DV Doppler become absent or reversed, fetal survival of longer than one week is unlikely 17 .
In late FGR, risk of terminal deterioration and unanticipated stillbirth is high, warranting closer surveillance. There are no evident Doppler changes in the precordial veins [pre-load to the fetal heart] and brain sparing may be the only observed Doppler sign of hypoxia. The new onset of Doppler abnormalities signals to expedite the birth 3,4 . Closer surveillance is required after 34 weeks because more fetal needs, less placental reserve and less time to adapt.

Planning birth
Gestational age should be the intervention threshold. Neonatal survival, more importantly neurologically intact survival, improves with gestational age. Planning birth is a decision taken after balancing the risks and benefits of birth compared to the continuation of pregnancy.
1. 24 to 26 weeks: Even with the best available care, the survival rate of growth-restricted neonates average less than 50% 18 . In the surviving babies, risks for major neonatal morbidity is high as 80%. The survival rates exceed 50% once the EFW exceeds 500 g and/or when 26 weeks have reached. Delivery at this age is more commonly due to maternal life-threatening conditions such as severe preeclampsia than FGR.
2. 26 to 28 weeks: Neonatal survival exceeds 50%. However, intact survival (survival without prematurity-related morbidity) at 26 to 27 weeks remains around 30%. Ductus venous Doppler abnormalities are not considered as an indication to birth until 28 weeks 3 while abnormal BPP (< 6/10) is an indication for delivery. Each day in utero increases neonatal survival by a median of 2% 3 .
3. 28 to 32 weeks: Neonatal survival exceeds 70% at 28 weeks and increases to more than 90% at 32 weeks. Fetal deterioration of venous Doppler parameters may be tolerated as long as DV flow is forward. Reversal of the DV a-wave before delivery reduces the chance of neonatal survival. Persistence of this abnormality beyond one week carries a significant risk for stillbirth. For this reason, the presence of a DV reversed a-wave is generally considered an indication of intervention from 28 weeks. However, birth before 30 weeks gestation still carries a significantly higher risk for adverse neurodevelopment at age two because of neonatal complications and their impact on motor development. Each day in utero increases neonatal survival by a median of 1%. These conclusions were based on TRUFFLE and GRIT trials and their follow up data 19,20,21,22 . Although there is no difference in neonatal mortality between immediate and deferred deliveries, due to a significant difference in morbidity in an immediate delivery group of early FGR, delaying at least up to 32 weeks is generally advocated based on these TRUFFLE and GRIT data 19,20,21,22 . 4. 32 to 34 weeks: Cerebral circulation gains an additional structural layer reducing cerebral haemorrhage risk during this period. This reduction has a measurable impact on motor development at age three. Administration of antenatal steroids has an added benefit in reducing respiratory morbidity and cerebral haemorrhage, and babies who have received steroids have improved survival. Evidence suggests that neurodevelopment is also improved by the administration of steroids 23 . Umbilical artery absent/reversed EDV is generally considered an indication for birth from 32 weeks onwards.

Neurodevelopmental outcomes in FGR
In early FGR, particularly before 28 weeks, gestational age at birth and birth weight are the most important predictors for poor neurodevelopment, advanced cardiovascular deterioration, and ICH leading to psychomotor retardation. Whereas in late FGR, gestational age at birth is a smaller contributor and other main predictors have yet to be defined. It mainly affects behavioral rather than motor development.

Cerebroplacental ratio and 'term FGR'
Many groups have suggested the CPR as a functional marker of failure to reach growth potential at or near term 26 . Low CPR, regardless of the fetal size, is independently associated with the need for operative birth for presumed fetal compromise and neonatal unit admission at term 26 . Rising evidence proposes that CPR as a marker of impaired fetal growth velocity and adverse pregnancy outcome, even in fetuses whose size is considered "appropriate for gestational age" using conventional fetal biometry 26,27 . Therefore it is wise to apply CPR at 40 weeks assessment of lowrisk pregnancies with fetal biometry to assess the suitability to wait a few more days without intervention.

Conclusion
Early FGR is associated with significant uteroplacental blood flow abnormalities and abnormal cardiovascular findings with evolving deterioration. In contrast, late FGR is not dominantly associated with placental vascular blood flow abnormalities and shows minimal signs of deterioration making it difficult to predict. EFW below the 10 th centile with co-existing UAD and/ or MCA Doppler flow and/or CPR abnormalities can identify a small fetus with pathological growth restriction requiring antenatal surveillance and planning delivery. Gestational age is the primary determinant as the threshold for intervention in planning delivery. Decision of timing of delivery is a balanced decision after considering fetal risks of continuing pregnancy and neonatal risks after delivery.

Author Declarations
Competing interests: The authors declare that they have no competing interests.

Funding:
No funding sources received.

Authors' Contributions:
First author, MP did the conception, design, data collection, analysis and writing/editing of the paper and he is the principal author for this work. Author AJ participated in writing and editing of the paper. Both authors approved the final manuscript to be published.