The cerebroplacental ratio: a useful marker but should it be a screening test?

The placenta and fetal brain are hemodynamically linked, as shown in animal models1-3 and humans4, 5. Nutrient and oxygen supply to the fetal brain depend, among other sources, on blood flow from the placenta through the umbilical vein4, 5. Placental insufficiency can be detected by increased resistance in the umbilical artery (UA), while the fetal brain response to hypoxia, known as the brain-sparing effect, can be observed using Doppler ultrasound, demonstrating widening of the middle cerebral artery (MCA) and blood flow redistribution. Both measurements have wide normal ranges, however, their value has been demonstrated in the monitoring of various fetal conditions, such as placental insufficiency and fetal growth restriction (FGR). The cerebroplacental ratio (CPR) was devised as a measure intended to detect incipient pathology earlier than can be detected by either of its constituent indices alone. With regard to late-onset FGR, CPR appears to be effective in the management of delivery timing. In recent years, various studies have investigated the use of CPR as a screening test in appropriate-for-gestational-age (AGA) fetuses, aiming to identify preclinical placental insufficiency and the consequent increased risk of fetal intolerance of labor. In this Opinion, we discuss the role of CPR as a useful Doppler parameter in fetal assessment that, nevertheless, currently falls short as a screening test. CPR is a Doppler-derived index that combines a placental Doppler measure (UA pulsatility index (PI)) and a fetal measure (MCA-PI) to assess fetal wellbeing. Its utility in monitoring various suspected obstetric complications is subject to some debate6, 7. The effectiveness of UA-PI assessment in the diagnosis and follow-up of placental insufficiency has been validated in many clinical settings8, but MCA-PI and CPR have shown varying results9-13. MCA Doppler assessment has been shown to be effective for monitoring fetuses with various conditions, including anemia14, 15. Its value in fetal anemia is based on decreased blood viscosity and increased fetal cardiac output which result in increased peak systolic velocity (PSV), detected on Doppler16. MCA Doppler measurement has been shown to provide evidence of fetal hypoxic distress and is effective in clinical management, interventional procedure timing and follow-up in the context of fetal anemia15. On the other hand, CPR was devised as an index that combines increased UA resistance, a marker of placental dysfunction and hypoxia, with increased diameter and decreased resistance in the fetal MCA to track the fetal response to hypoxia. CPR measurement is thought to be especially relevant to those cases with UA-PI and MCA-PI preserved within the normal range that seem to trend toward increased placental resistance and decreased MCA resistance, which may be suggestive of the early stages of emerging brain sparing; 'low' CPR is applied as an indicator of fetal brain sparing. There is some debate as to whether this cerebral autoregulation, involving preferential redistribution of blood flow to the brain, heart and adrenal glands, is a protective physiological mechanism or a salvage pathway in response to hypoxia6, 17-20. CPR is classified as 'low' around various cut-off values (≤ 1.0, 1.08 or 1.1; < 5th, 10th or 20th centile; < 0.6765 multiples of the median)6, 21-33, which provide a snapshot of the relative magnitudes of the two component parameters as a binary measure. Reference ranges for MCA-PI and other fetal Doppler measures, including CPR, show considerable variability34, 35. There is a lack of consensus regarding the optimal reference range to be used and what constitutes 'low' CPR9, 35-38. The International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) practice guidelines35 also stress the importance of repeated measurements of the MCA-PI. Cross-sectional39 and longitudinal40 CPR studies have found a 'U-shaped' variation in mean CPR over the course of gestation, with a wide distribution among normal fetuses in each week of pregnancy, as well as a steady difference in mean CPR between male and female fetuses40. UA-PI shows a linear decrease with increasing gestational age in normal fetuses41. Hence, the U-shaped variation in mean CPR would appear to stem from normal variation in MCA Doppler measurement. It is recognized that low estimated fetal weight (EFW) alone, while an important measure of fetal wellbeing, is a weak predictor of fetal compromise, since it classifies up to 10% of all fetuses as potentially small-for-gestational age (SGA), by definition. Most SGA fetuses are constitutionally small but healthy, and will be born without complications, while up to two-thirds of compromised neonates were not suspected to be SGA or considered at risk of peripartum complications. Therefore, in the case of small fetuses, there is a need to differentiate between constitutionally small but healthy fetuses and those with FGR. Among apparently AGA fetuses, it is especially important to identify those with mild placental dysfunction which may be missed by UA-PI or other markers. As per the Delphi consensus42, FGR is divided according to the severity and timing of placental dysfunction into early-onset FGR, diagnosed at < 32 weeks' gestation, and late-onset FGR, diagnosed at ≥ 32 weeks' gestation42-44. The American College of Obstetricians and Gynecologists (ACOG) defines FGR more broadly, as fetuses with abdominal circumference (AC) or EFW < 10th centile for gestational age45. Early-onset FGR (in the absence of congenital anomaly) is defined by ISUOG as AC/EFW < 3rd centile or absent UA end-diastolic flow, or AC/EFW < 10th centile combined with uterine artery PI (UtA-PI) > 95th centile and/or UA-PI > 95th centile44. Ductus venosus a-wave and UA Doppler measures have been shown to provide reliable evidence of fetal distress in early-onset FGR, which can inform clinical management44. MCA-PI and, by extension, CPR assessment are not recommended in this population6, 40, 46, 47. ISUOG guidelines on early-onset FGR do not recommend CPR assessment as either a diagnostic tool or as a part of management protocols44. Late-onset FGR, appearing at ≥ 32 weeks' gestation, is defined by ISUOG as AC/EFW < 3rd centile or at least two of the following: AC/EFW < 10th centile; AC/EFW crossing centiles > 2 quartiles on growth centiles; CPR < 5th centile or UA-PI > 95th centile44. MCA-PI Doppler assessment is used in late-onset FGR, with low MCA-PI usually resulting from increased diastolic flow. Low MCA-PI and CPR were shown to be more frequent in late-onset FGR48 and to be associated with adverse outcomes, including Cesarean delivery for fetal distress and admission to the neonatal intensive care unit (NICU)24. The ISUOG guidelines recommend using MCA-PI and its ratio to UA-PI (i.e. CPR) to monitor late-onset FGR; however, CPR is not endorsed in all protocols of FGR evaluation, such as the ACOG/Society for Maternal–Fetal Medicine (SMFM) practice bulletin45. While the ISUOG protocol for late-onset FGR includes CPR as part of the diagnostic criteria, it is only included in management decision-making after 38 weeks' gestation44. Screening tests are performed on asymptomatic individuals to identify diseases or complications before symptoms appear, allowing for early intervention and management aimed at improving health outcome. Indeed, screening is not recommended unless a clear next step (e.g. further testing, intervention) is available and known to be beneficial. Screening tests must be safe, cost-effective and widely accessible, and they should have high sensitivity and specificity to minimize false positives and false negatives. Prior to recommending the adoption of a test as a screening test, it is imperative to consider both the cost–benefit and risk–benefit analyses, to ensure the test is effective and efficient, reduces potential harm caused by the condition being tested for and minimizes iatrogenic harm caused by any subsequent intervention. There is a lack of consensus regarding the sensitivity and specificity of CPR assessment in AGA fetuses and what result would indicate that an asymptomatic mother–AGA-fetus dyad is at high risk of intrapartum complications (i.e. 'screen positive'), and what the next steps would be in such a case9-12. The cost of screening includes not only the expense of the test itself, the infrastructure needed and the healthcare personnel required to oversee it, but also the time it will take away from patient care and patient interaction with caregivers. Screening often leads to additional consultations, diagnostic tests and treatments, which should be factored into the overall cost. In addition, screening procedures may come with risks of physical or psychological harm. In the case of CPR, potential harm may include risks to the fetus from iatrogenic early delivery, or for the parents, anxiety caused by the test result. The risk of overdiagnosis in asymptomatic use of CPR evaluation could have a negative impact on parents and infants. CPR would be hard-pressed to achieve the aforementioned benchmarks of a screening test. Firstly, MCA-PI measurement is difficult to obtain in the third trimester, certainly at or near term when the fetal head is engaged, and shows wide variation between centers, over the course of gestation and between male and female fetuses34, 35, 39, 40, 44. Variation may stem from the technical difficulty in obtaining measurements, interobserver differences, changes in Doppler waveforms caused by transducer pressure on the fetal head, fetal activity and maternal factors, such as body habitus and position. Owing to the known inter- and intraobserver variability in MCA-PI measurement, repeated measures are recommended by the ISUOG guidelines35 and others39. Any abnormal value measured in clinic should be re-examined to exclude maternal factors and technical artifacts, and be considered within the clinical context, considering parameters such as background risk factors, EFW, fetal growth trajectory and UtA Doppler. Since MCA-PI measurement is integral to CPR but cumbersome in clinical practice, CPR is not an optimal candidate for a screening test. Furthermore, the lack of standardization of Doppler measures in obstetric ultrasound36-38, 44, including CPR, undermines attempts to establish real-world sensitivity and specificity for CPR in screening-level assessment for the risk of intrapartum complications in apparently healthy AGA fetuses. Unlike UA-PI, which is validated in the clinical setting8, MCA-PI, CPR and UtA-PI have produced inconsistent results or added no value in the prediction of adverse perinatal outcomes beyond that of UA-PI9, 12, 22, 23. In the first study related to CPR in AGA fetuses at term23, measurements were performed immediately prior to delivery, which is relevant to fetal wellbeing at the time of delivery. In another study21, the authors show an association between CPR and neonatal outcome when measurement was performed in closer proximity to delivery and later in gestation (≥ 38 weeks), but CPR was not predictive of neonatal outcome when performed weeks before delivery and earlier in gestation (< 38 weeks). Use of the CPR for the prediction of perinatal complications has been reviewed by several groups using various approaches9-12, although authors note the heterogeneity of timing of testing prior to delivery and the cut-off used to define 'low' CPR. Flanagan et al.11 performed a pooled analysis of individual patient data from three cohort studies to determine the prognostic value of various Doppler measures in the prediction of adverse perinatal outcomes. The researchers devised six models to assess the prognostic value of Doppler measures using continuous PI values at various gestational ages and birth-weight (BW) centiles. They showed that, in preterm fetuses (< 37 weeks' gestation), EFW enhanced the area under the curve (AUC) to 0.771. When successively adding UA-PI and MCA-PI to the model, AUC increased to 0.803. However, when replacing UA-PI and MCA-PI with CPR, the AUC decreased to 0.785. In cases delivered at term, when using the base model, which included maternal age, gestational age, parity, maternal smoking and body mass index, the AUC was 0.678; the addition of EFW and Doppler parameters to the model did not significantly impact the result11. The investigators also divided the pooled individual patient data into BW cohorts and found that when BW was < 10th centile, the base model and the model including EFW had AUCs of 0.729 and 0.745, respectively. Adding UA-PI to the model increased the AUC to 0.770 and further addition of MCA-PI and UtA-PI increased the AUC to 0.792, while replacing UA-PI and MCA-PI with CPR gave an AUC of 0.775. In the middle BW subgroup (10th–25th centile) the base model AUC was 0.674. The addition of EFW and Doppler measures did not increase the AUC and the addition of CPR to the model gave an AUC of 0.675. The same lack of improvement was observed in those with BW > 25th centile11. Screening tests should balance the risk of unnecessary and iatrogenic intervention with risk of harm from the screened-for condition. In the case of CPR assessment in low-risk fetuses near term, there is currently no next step for screen-positive cases that has shown any benefit. It is incumbent on professional organizations to ensure that screening tests are only recommended when they provide a clear net benefit to the population. CPR seems to currently fall short as a screening test in low-risk, term pregnancies. Randomized controlled trials (RCTs) are considered to set the standard for evidence-based medicine, sitting in second place on the pyramid of evidence underneath meta-analyses and systematic reviews. Pragmatic RCTs aim to evaluate the effectiveness of interventions in real-world clinical settings, rather than under controlled conditions typical of traditional RCTs. This design choice reflects the variability and challenges encountered in real-world settings and aims to strengthen the validity of trial findings for routine clinical use. Many CPR studies in low-risk fetal populations have concluded with calls for RCTs to be conducted to provide stronger evidence for its use. There are inherent problems with pragmatic and traditional RCTs and how they are perceived in the medical community. There is a tendency to view the results of RCTs as outweighing all other types of evidence and to accept their findings unreservedly. A case in point is the trial that led to accelerated approval of 17-hydroxyprogesterone caproate for recurrent preterm birth prevention49 and the later abandonment of this intervention when it was proven to be ineffective50. Part of open academic discourse is the necessity to debate the results of clinical trials and their implications in clinical care. It is important to recognize that there are drawbacks, even to well-designed pragmatic or traditional RCTs, and that they can be subject to misinterpretation51, 52. Composite outcomes may generate results that appear to be statistically significant, despite the primary outcomes of interest failing to reach statistical significance or to demonstrate any benefits from the intervention being tested51, 52. A recent pragmatic RCT53 sought to evaluate whether adding CPR to routine third-trimester fetal growth assessment would reduce the rate of poor neonatal outcomes in a cohort of presumed low-risk fetuses, compared with standard care. Pragmatic trials intentionally embrace variability to assess the intervention's true effectiveness across different settings; RATIO37 was a large multinational trial that recruited around 11 000 participants53. The authors assessed CPR between 36 and 38 weeks' gestation, requiring the observation of three or more similar waveforms consecutively but making no mention of repeated measurements53. The CPR results were either revealed to or concealed from delivery room caregivers and the obstetric and neonatal outcomes of the two groups were compared. The concealed group included a slightly higher proportion of fetuses with CPR < 5th centile (6.8% vs 5.6%). There was no difference in gestational age at delivery or birth weight between the groups, nor in the rates of labor induction or elective Cesarean delivery. Caregiver knowledge of CPR measurements seemed to be associated with an increase in instrumental delivery but the investigators found no difference in the rate of instrumental or Cesarean delivery for non-reassuring fetal status53. There were similar, very low rates (< 1%) of the quantified severe outcomes in the two groups: perinatal death, intraventricular hemorrhage, periventricular leukomalacia, hypoxic ischemic encephalopathy, necrotizing enterocolitis, renal failure and cardiac failure. There was no significant difference in the rate of NICU admission (10.6% vs 10.4%, P = 0.748). The only secondary outcome reported to reach statistical significance was a reduction in NICU admission for ≥ 10 days in the group with revealed CPR results. Unfortunately, no details regarding these infants or their hospitalization, such as gestational age at delivery, birth weight, indication for NICU admission or the median, range or interquartile range of their length of stay were provided. This would have allowed for comparison across the diverse patient populations and clinical environments included in this pragmatic trial. When only fetuses with CPR < 5th centile in the concealed and revealed groups were compared, no significant differences were found, including in the rate of SGA53. The association of FGR with low CPR and risk of peripartum complications is well known. When only fetuses in the concealed group with CPR ≥ 5th centile vs < 5th centile were compared, only the rate of SGA differed significantly53. In both of these subgroup comparisons, NICU admission for ≥ 10 days was again reported without further details. The authors concluded with a call for further studies to determine whether CPR assessment is effective in all pregnancies or only selected groups, but still recommend universal CPR screening in the third trimester, even though CPR assessment failed to show any reduction in perinatal mortality, severe/mild neurological outcomes or any other outcome, other than NICU admission for ≥ 10 days. A policy of incorporating CPR assessment into routine, low-risk EFW scans could lead to an increase in the unnecessary induction of normal fetuses. This exposes mothers and fetuses to the known risks of induction, as well as the loss of the important final weeks of in-utero brain development. The authors of the aforementioned RCT state that 342 screening tests would be required to prevent one event of severe neonatal morbidity, and at a screen-positive rate of 5%, 17 pregnancies would be recommended for planned delivery, and in their population, would face a slightly higher risk of operative vaginal delivery53. Early-term delivery itself has been shown to be associated with various neonatal and pediatric adverse outcomes. The authors did not address how this screening would be applied in line with the '39-week rule'54, 55. CPR assessment and its component measures have demonstrated benefits in the monitoring of some subgroups of high-risk fetuses. CPR is not included in the diagnostic criteria or management protocols for early-onset FGR, although it is useful after 38 weeks' gestation for decision-making regarding delivery in late-onset FGR. CPR is a cumbersome measurement that has not yet been shown to have any added value in a screening population and has real potential for causing a negative impact by creating unnecessary stress for patients, as well as unnecessary interventions and a strain on resources, by diverting time from other more meaningful contact with parents. A pragmatic RCT published in The Lancet has the potential to affect global clinical practice in this field. However, based on the current evidence, including the findings of this pragmatic RCT, it does not currently appear justified to carry out general peripartum screening by means of CPR measurement in AGA fetuses. Data sharing is not applicable to this article as no new data were created or analyzed in this study.