Considering evidence in the management of fetal growth restriction

Despite a significant body of evidence regarding the clinical picture of fetal growth restriction (FGR) due to placental dysfunction, management guidance remains non-unified across professional societies1. In the absence of effective treatment, established placental dysfunction places the fetus at risk for progressive nutrient deprivation, hypoxemia, acidemia, organ damage and/or death2. Prenatal management still focuses primarily on weighing the existing and future risks for the undelivered fetus against those of the neonate following delivery. Therefore, an optimal surveillance strategy has to reflect accurately the current fetal status as well as provide a safety net against unanticipated fetal demise for patients not yet meeting thresholds for delivery. The underlying placental pathology is responsible for the variation in FGR phenotype as gestational age advances2, 3. Second-trimester FGR is characteristically associated with significant villous malperfusion that leads to elevated umbilical artery (UA) blood-flow resistance, progressing to absent end-diastolic velocity (AEDV) and even reversed end-diastolic velocity (REDV)4, 5. Additional signs of placental dysfunction, such as fetal growth arrest, oligohydramnios, brain-sparing, abnormal venous Doppler, heart-rate decelerations and an abnormal biophysical profile (BPP), are observed in over 80% of FGR cases delivered by 32 weeks' gestation6, 7. Fetal demise at delivery occurs in 0–4% of cases8. While third-trimester placental pathology still impairs villous gas diffusion, perfusion is less affected and UA blood flow may be normal. After 32 weeks' gestation, the cumulative frequency for significantly abnormal UA or ductus venosus Doppler are only 5% and 0.1%, respectively5, 8, 9. Approximately 20% of fetuses with normal UA Doppler show middle cerebral artery (MCA) dilation. Collectively, decreasing amniotic fluid, brain-sparing, loss of heart-rate reactivity and abnormal biophysical parameters are seen in approximately 40% of cases5, 9. For patients remaining undelivered, the prospective stillbirth rate of 11/10 000 doubles weekly after 37 weeks' gestation4, 5, potentially resulting in stillbirth rates as high as 8% at the time of birth. For delivered patients, gestational age-specific risks are significantly higher for growth-restricted than for appropriately grown neonates. Fifty percent of growth-restricted fetuses delivered at 26 weeks' gestation are expected to survive, especially if their estimated fetal weight is > 500 g. For delivery between 26 and 30 weeks, there is an accelerating decline in neonatal mortality, morbidity and neurodevelopmental delay, with a residual need for intensive care admission until 38 weeks' gestation, with negligible neonatal morbidity5, 10 (Figure 1). It is only from 28 weeks onward that the degree of compromise at birth impacts neonatal complications independent of prematurity5. In view of the inconsistency in viability before 26 weeks, the emphasis of management at this age lies on maternal health, as gestational hypertensive disease accompanies at least 25% of cases, and consideration of fetal status is individualized. The required certainty for the decision to deliver for fetal indications decreases with advancing gestation, as the corresponding neonatal risks decline. Fortunately, early-onset FGR presents with a multitude of surveillance abnormalities that can guide delivery decisions. As the emphasis shifts towards lower delivery thresholds between 34 and 38 weeks, the clinical management becomes challenged by the paucity of warning signs of fetal compromise4, 5, 11. Because of the changing FGR phenotype as gestation progresses and the physiologic basis of surveillance tests, the challenges of preventing iatrogenic delivery and stillbirths cannot be provided by a single surveillance modality. Abnormal UA end-diastolic velocity is a consequence of villous vascular malperfusion, and the relationship with fetal hypoxemia is by virtue of the associated impairment in transplacental gas exchange2, 11. MCA end-diastolic velocity increases in response to hypoxemia and the accompanying increase in fetal blood pressure. The Doppler cerebroplacental ratio (CPR) of MCA-PI to UA-PI is influenced by vascular changes in either vessel increasing the sensitivity for hypoxemia. Atrial systolic forward velocities in the ductus venosus decrease with vessel dilation, myocardial dysfunction and escalating UA blood-flow resistance which occur after prolonged severe placental dysfunction associated with fetal acidemia2, 5, 11. Amniotic fluid volume decreases with impaired uteroplacental blood flow and, therefore, relates indirectly to the fetal pH and, by virtue of the increased risk for cord accidents, to increased likelihood of stillbirth2, 11. Fetal biophysical variables (heart rate (FHR), breathing movements, gross body movement and tone) are controlled by centers in the fetal brain that are directly sensitive to varying degrees of hypoxia, irrespective of its underlying cause. Heart-rate reactivity and breathing movements cease below a pH of 7.20, and fetal movement and tone below a pH of 7.102, 4, 5. Since observation of individual biophysical variables also relates to fetal behavior and maturation, their combination enhances the accuracy of prediction of fetal status by accounting for physiologic variation. FHR reactivity is most sensitive to fetal oxygenation and is therefore universal in fetal surveillance. Computerized heart-rate analysis (computerized cardiotocography (cCTG)) extracts key heart-rate variables, including short-term variation (STV) during fetal activity, which are then analyzed accounting for gestational age by the Dawes–Redman criteria. The modified BPP combines visual FHR reactivity assessment with the amniotic fluid index. The multicomponent BPP score combines four variables related to current fetal status (FHR, tone, movement, breathing) with amniotic fluid volume or placental grading as parameters predictive of future placental function5, 12. The principal difference between Doppler and biophysical surveillance modalities lies in their correlation with hypoxia and the clinical progression of placental dysfunction. Doppler parameters reflect cardiovascular responses that directly predict impending deterioration but are only indirectly associated with acid–base status2, 5, 11. In contrast, biophysical variables are gradually lost with worsening hypoxia, independent of gestational age. Because biophysical parameters monitor the impact rather than the severity of placental disease, even combinations of tests cannot predict compromise accurately2, 5, 11. These critical physiologic differences dictate the optimal FGR surveillance strategy. Every FGR surveillance visit has to accurately reflect current fetal status as well as inform about the follow-up interval that is required to pre-empt significant deterioration or even stillbirth. This is accomplished only by concurrent consideration of Doppler and biophysical parameters. An abnormal BPP most accurately predicts prelabor acidemia (Figure 2) and, together with FHR decelerations, is a delivery indication at any gestational age (Figure 3). In early-onset FGR, all other fetal delivery indications require consideration of gestational age. The Trial of Randomized Umbilical and Fetal Flow in Europe (TRUFFLE) study6 established an absent ductus venosus a-wave or a low cCTG FHR STV as delivery criteria from 26 weeks onward. UA Doppler parameters should be considered as delivery indicators only after 30 weeks, as the pooled stillbirth rate of 20% for REDV and 6% for AEDV exceed the neonatal mortality rate from 30 and 32 weeks, respectively13. In TRUFFLE, 11% of patients randomized to severe ductus venosus Doppler abnormality delivered for an absent or reversed a-wave, while 36% delivered for FHR decelerations and 23% for STV safety criteria10. In this setting, 24% would be expected to have an abnormal BPP. Importantly, in approximately 30% of patients, delivery was delayed after 32 weeks. For late-onset FGR, optimal delivery triggers remain to be determined, and observational cohorts indicate that 32% are delivered for unspecified FHR or Doppler parameters, 28% for growth deceleration and 10% for maternal indications14, and 13% of these would be expected to have an abnormal BPP7. Surveillance frequency in FGR has not been evaluated in a dedicated study. As biophysical parameters add little guidance in this regard, consideration of cardiovascular parameters is necessary. Oligohydramnios universally increases risk for FHR decelerations, and increasing surveillance frequency should be considered irrespective of gestational age. With advancing gestational age, prediction of deterioration and stillbirth is provided by different vessels. In early-onset FGR, deterioration is best predicted by UA Doppler and stillbirth by ductus venosus Doppler. In late-onset FGR, deterioration is predicted by CPR and stillbirth by MCA brain-sparing8 (Figure 3). Despite these important recognized benefits of the array of aforementioned surveillance tests, only traditional FHR testing and UA Doppler are recommended uniformly by professional societies, while additional recommendations vary according to geography. In addition to the complexities of the disease itself, these variations in recommendations are likely to create additional challenges for managing physicians (Table 1). High FHR sensitivity for reassuring fetal status FHR abnormality constitutes majority of delivery indications across gestational age UA Doppler guides surveillance frequency in early-onset FGR by capturing accelerating placental dysfunction (1) High rate of non-reactive NST requires secondary tests (2) Does not provide guidance on surveillance interval in majority of late-onset FGR (3) Delivery timing based on UA Doppler will lead to iatrogenic prematurity in early FGR (4) With markedly abnormal UA Doppler there are no further warning signs for stillbirth Allows calculation of Doppler CPR Together with CPR, provides guidance on surveillance frequency in late FGR Provides evidence-based guidance on timing of delivery from 26 weeks onwards Provides anticipation of stillbirth in early-onset FGR Examination expertise is not uniformly available Does not address disadvantages (1) and (2) of UA/NST monitoring Provides a high negative prediction of stillbirth if combined with NST (modified BPP) Gives guidance on surveillance frequency across gestational age Provides risk-based delivery indication independent of gestational age Provides comprehensive secondary test if NST is non-reactive Provides high negative prediction of stillbirth Examination expertise is not uniformly available Does not provide guidance on surveillance frequency Does not address disadvantages (2) and (4) of UA/NST monitoring Provides evidence- and risk-based delivery guidance from 26 weeks onward Enhances accuracy of FHR analysis Not uniformly available Does not provide guidance on surveillance frequency Does not address disadvantages (2) and (4) of UA/NST monitoring Without a structured, consistent management approach tailored to the specific clinical features of FGR, iatrogenic prematurity and stillbirth rates are likely to increase – a consequence that is clouded by research that does not specifically distinguish between prenatal and neonatal contributors to perinatal outcome. Management will become unified only after universal agreement on the key metrics of prenatal management, adherence to core outcome sets for research reporting15 and analysis of the contribution of suboptimal surveillance intervals to stillbirth. With ever-increasing global information exchange, guidelines should provide explanation of discrepancies in their recommendation with those of other societies to avoid confusion of patients and physicians alike. Arguments favoring one surveillance modality over another are unhelpful, as their optimal integration is the only approach that will advance our management goals from survival to favorable long-term outcome and quality of life.