Diagnosis and management of fetal growth restriction: the <scp>SMFM</scp> guideline and comparison with the <scp>ISUOG</scp> guideline

Abstract

Fetal growth restriction (FGR), broadly defined as a fetus not reaching its growth potential, is caused by maternal, fetal and placental conditions that contribute to suboptimal placental perfusion, fetal nutrition and, in some cases, oxygenation1, 2. FGR is associated with significant short- and long-term morbidity and mortality1-3. The application of a structured guideline for the clinical management of pregnancies with FGR has been shown to enhance standardization of care and improve outcome4. The Society for Maternal–Fetal Medicine (SMFM) has recently released a revised and comprehensive guideline on the prenatal diagnosis and management of FGR5. This Opinion is intended to highlight important aspects of the SMFM FGR guideline and provide the rationale for discrepant elements when compared with the FGR guideline of the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG)6. The SMFM defines FGR as an ultrasonographic estimated fetal weight (EFW) or abdominal circumference below the 10th percentile for gestational age5. Despite the fact that a significant number of constitutionally small fetuses are included below the 10th percentile threshold for FGR, large population-based studies have consistently shown increased perinatal morbidity and mortality below this weight threshold, with pregnancy risks comparable to, or higher than, those associated with other high-risk pregnancy conditions, with established fetal surveillance patterns. In a retrospective cohort study including all singleton neonates born in the USA in 2005, the risk for intrauterine fetal demise in those with birth weight < 10th percentile was 5-fold higher than in those with birth weight ≥ 10th percentile7. A meta-analysis of 28 studies found that newborns with weight below the 10th percentile had significantly lower standardized neurodevelopmental scores8, and a large long-term epidemiologic study found that term newborns with birth weight < 10th percentile had an increased risk for adult-onset diabetes when adjusted for body mass index and parental history of diabetes mellitus9. Finally, a study evaluating the placentas of singleton fetuses with EFW < 10th percentile in the third trimester and normal umbilical artery (UA) Doppler, that were delivered after 34 weeks' gestation, found that placental histopathologic signs of underperfusion were present in 66.7% of cases and correlated with a higher incidence of emergency Cesarean section and neonatal acidosis10. Interestingly, middle cerebral artery (MCA) and cerebroplacental ratio (CPR) Doppler indices did not predict pathologic markers of underperfusion in that study10. In view of these findings, we believe that the 10th percentile EFW threshold seems to be the most suitable cut-off for the diagnosis of FGR and is currently the most commonly applied threshold for FGR diagnosis in several national guidelines11. In an attempt to distinguish between FGR and constitutionally small fetuses, consensus-based definitions for both early- and late-onset FGR were established through a Delphi procedure12 and recently adopted by the ISUOG FGR guideline6. The strength of the Delphi procedure is in creating consensus between a panel of experts, through a series of sequential rounds of questions, on topics that cannot be answered by clinical research13, 14. Since FGR research is evolving, the Delphi approach carries the risk of introducing definition parameters into clinical practice based on opinions, without the benefit of scientific vigor. For instance, uterine artery Doppler was included in the Delphi definition as a contributory parameter for the diagnosis of early FGR despite its low sensitivity (25.8%) in predicting composite adverse pregnancy outcomes15. Although there is an association in late FGR between abnormal CPR and short-term adverse outcome, there is no strong evidence to date on whether incorporating CPR in clinical management actually improves outcome16 and, therefore, we feel that its inclusion in the definition of late FGR should await the results of the planned randomized trial on this subject16. Finally, the proposed Delphi definition of FGR will increase testing utilization, as uterine artery or MCA Doppler testing will be required in about 10% of pregnancies. In this issue of the Journal, Roeckner and colleagues compared the performance of the SMFM and ISUOG criteria for the diagnosis of FGR in predicting a small-for-gestational-age (SGA; birth weight < 10th percentile) neonate and neonatal morbidity17. The SMFM criteria had higher sensitivity than the ISUOG criteria for the prediction of neonatal SGA (54.7% vs 28.8%) but had a higher false-positive rate (6.7% vs 1.6%). The positive predictive value for the prediction of neonatal morbidity was poor for both definitions (15.3% vs 25.5%), which likely reflects, at least in part, the rarity of adverse neonatal outcomes in pregnancies diagnosed with FGR that undergo antenatal surveillance and appropriate management17. Based on these data, we believe that the simplicity of the application of the SMFM criteria offers a more straightforward and pragmatic approach to the diagnosis of FGR, that also has the ability to identify more at-risk pregnancies17. There is consensus among FGR guidelines on the importance of UA Doppler in fetal surveillance and timing of delivery11. However, data to inform the frequency of UA Doppler evaluation are limited. Based on the SMFM guideline, upon diagnosis of FGR initial evaluation every 1–2 weeks is recommended, with subsequent evaluation every 2–4 weeks if the UA blood flow remains normal5. Findings indicative of decreased end-diastolic flow (EDF) in the UA should be evaluated weekly. When absent EDF in the UA is noted, assessment 2–3 times per week is recommended and evidence of deterioration to reversed EDF should warrant further escalation of care and surveillance in preparation for delivery5. Cardiotocography (CTG) is well-accepted as a primary surveillance tool in high-risk pregnancies5. In many management schemes, CTG has been cited as a standard monitoring tool, despite the lack of rigorous studies proving its efficacy18. Abnormal CTG with loss of fetal heart rate variability has been associated with acidosis and hypoxemia, and the presence of spontaneous repetitive decelerations is a trigger for delivery in viable fetuses with FGR5, 6, 19, 20. It is important to note that abnormal CTG in FGR may be present in the absence of UA Doppler abnormalities and may represent an alternate pathway of fetal deterioration. The SMFM does not recommend Doppler assessment of the MCA or ductus venosus (DV) in the clinical management of FGR, which is a significant difference from the ISUOG guidelines5, 6. Evidence supporting these recommendations is presented in the following sections. In the setting of FGR, MCA Doppler provides information about cerebral vasodilation, an early adaptive mechanism of fetal hypoxemia21-23. However, there is currently limited evidence to suggest that its incorporation into clinical decision-making in early FGR actually improves outcome. In a meta-analysis of 35 studies on the subject, abnormal MCA Doppler had limited positive predictive accuracy for perinatal mortality and adverse perinatal outcome24. In a secondary analysis of data on early FGR from the TRUFFLE (Trial of Randomized Umbilical and Fetal Flow in Europe) study, MCA Doppler was found to have a modest association with neonatal and 2-year infant outcome and did not add useful information beyond UA and DV Doppler for optimizing the timing of delivery25. As such, there is currently no compelling evidence to support the use of MCA Doppler in fetal surveillance or in delivery timing in early FGR2, 24, 25. In late FGR, studies have demonstrated that 15% to 20% of fetuses with normal UA Doppler have MCA Doppler findings of cerebral vasodilation26. CPR has also been studied for its utility in predicting adverse outcomes and guiding delivery timing in late FGR27-31. Recently, a prospective multicenter observational feasibility study, involving 856 pregnancies with late FGR, was undertaken as part of the design process for the TRUFFLE-2 randomized trial for determining arterial Doppler thresholds that are most strongly associated with adverse outcome and optimal timing for delivery16. The study showed that the first Doppler observation of MCA pulsatility index < 5th percentile and umbilicocerebral ratio Z-score above gestational-age-specific thresholds had the highest relative risks for composite adverse outcome, although gestational age at delivery and birth-weight Z-score showed a stronger association16. The authors concluded that it is still unclear whether cerebral redistribution is a marker of severity of FGR or an independent risk factor for adverse outcome, and that its usefulness in clinical management can be answered only in a randomized trial16. We therefore believe that the incorporation of cerebral Doppler in the management of late FGR should await further evidence from randomized trials5. In the setting of severe early FGR, spectral Doppler abnormalities of the DV reflect an advanced stage of fetal compromise and are associated with significant increase in perinatal morbidity and mortality32-36. The role of DV in the clinical management of pregnancies with early FGR was evaluated prospectively in the TRUFFLE trial which compared the efficacy of DV Doppler with that of computer-generated fetal heart rate short-term variation (cSTV) in fetal monitoring and timing of delivery4. Infant survival without neurological impairment at 2 years of age was significantly higher in the group delivered according to late DV abnormalities (95%) when compared with the group delivered according to cSTV (85%)4. The authors concluded that incorporating DV Doppler in the management of early FGR can guide the decision for delivery and possibly reduce long-term neurological sequelae4. The ISUOG guidelines for FGR include DV Doppler abnormalities as a trigger for delivery in pregnancies with early FGR6. Despite the evidence that DV Doppler accurately predicts perinatal morbidity and mortality in early FGR, its use in the clinical management of FGR is not recommended in the SMFM FGR guidelines5. This deliberate decision was based on the following points. The TRUFFLE trial was designed to compare delivery triggers between cSTV and DV Doppler abnormalities and, thus, caution should be exercised in extrapolating these findings to clinical settings that do not utilize cSTV, such as in the USA, but rely on visual interpretation of CTG. The safety-net criteria in the TRUFFLE trial that mandated delivery irrespective of the randomization arm, and which accounted for the delivery of 33% of pregnancies in the late-DV group, included cSTV cut-offs that are difficult, if not impossible, to identify on visual interpretation of CTG4. Indeed, in an editorial presenting the key messages from the TRUFFLE trial, the authors recommended against replacing visual interpretation of CTG with cSTV in settings in which cSTV is not available37. In addition, they stated that, if the results of the TRUFFLE trial are implemented in guidelines or local protocols, the safety-net criteria (which included cSTV) should be an integral part and that the TRUFFLE trial results are only generalizable in settings in which cSTV is available. Unlike the UA, DV waveform abnormalities, including absent or reversed a-wave, may result from cardiac or vascular abnormalities in the absence of placental disease or fetal hypoxemia, and, thus, may partake in the inadvertent clinical decision to deliver a premature fetus, especially in settings that lack Doppler expertise. It is also important to note that the contribution of DV Doppler in the clinical management of early FGR is somewhat limited, as absent or reversed a-wave in the DV is encountered quite infrequently in the clinical setting. In a recent retrospective cohort study from two tertiary referral units, including 132 singleton growth-restricted fetuses with absent or reversed EDF in the UA, absent or reversed a-wave velocities in the DV were noted in only 15/132 (11%) of fetuses38. Indeed, in the TRUFFLE trial, delivery decision guided by DV Doppler abnormalities accounted for only about 10% of pregnancies allocated to the late-DV group, as most pregnancies were delivered due to safety-net criteria or other fetal or maternal indications39. In view of this limited role of DV in clinical decision-making in early FGR and the lack of prospective trials elucidating the role of DV Doppler in the setting of visual interpretation of CTG, there is currently insufficient evidence to recommend DV Doppler as part of a national guideline for clinical management of FGR in the USA5. With the understanding of limited generalization of the existing data on this subject, maternal–fetal medicine practices that wish to incorporate DV in the clinical management of early FGR may consider restricting its application to fetuses with absent or reversed EDF in the UA, in order to refine timing of delivery in the gestational-age window between 26 and 30 weeks of gestation. Rigorously developed guidelines in medicine have the power to simplify the complexity of scientific research and develop recommendations that can potentially enhance health quality and outcome. In particular, the development of a FGR guideline that reflects a critical appraisal of the existing evidence is fundamental to enhancing clinician and patient decision-making and in standardization of care. This is of utmost importance given the vast number of scientific publications on the subject and the impact of FGR on population health. In developing the SMFM FGR guideline, we have strived to maintain simplicity and clarity in order to enhance compliance and health equity. We have also ensured that the recommendations are applicable and relevant to existing practice patterns and that major decisions, such as delivery triggers, are based on solid evidence that actually impacts perinatal outcome rather than mere associations. Finally, we commend our colleagues, the authors of the ISUOG FGR guideline, for their massive contribution to the FGR literature and their continued quest for discovery to help answer critical questions on this subject. Indeed, we are all beneficiaries of these scientific discoveries as the FGR guidelines will evolve over time to reflect the current state of science on this subject.

Topics & Concepts