The value of introducing cerebroplacental ratio (CPR) versus umbilical artery (UA) Doppler alone for the prediction of neonatal small for gestational age (SGA) and short-term adverse outcomes

Abstract

Objective

To compare the role of UA Doppler versus CPR in the prediction of neonatal SGA and short term adverse neonatal outcome in a high-risk population.

Study design

We conducted a prospective study on women referred for fetal growth ultrasounds between 26 and 36 weeks of gestation and with an EFW < 20^th percentile by Hadlock standard. UA and middle cerebral artery (MCA) Doppler assessments were performed. Abnormal UA doppler was defined as: pulsatility index (PI) above the 95th percentile and absent or reverse end diastolic flow. The CPR, calculated as a ratio of the MCA PI by the UA PI, was defined as low if < 1.08. The primary outcome was the sensitivity and specificity of the two Doppler assessments to predict neonatal SGA, defined as birthweight < 10^th percentile by using Alexander curves. The secondary outcomes included umbilical cord arterial pH < 7.10, APGARS at 5 minutes < 7, NICU admission, respiratory distress syndrome (RDS), hypoglycemia or a composite including any of these secondary outcomes. Chi square was performed for statistical analysis.

Results

Of 199 women meeting inclusion criteria, 94 (47.2%) had SGA and 68 (34.2%) had a composite adverse outcome. A total of 7 pregnancies with FGR had a low CPR. Abnormal UA Doppler showed a better sensitivity for predicting SGA and adverse neonatal outcomes, with comparable specificity to low CPR. The area under the ROC curve (AUC) using abnormal UA Doppler for predicting SGA was 0.54, 95% CI 0.50–0.58; and 0.51, 95% CI 0.48–0. 53 for low CPR. The AUC for predicting a composite adverse neonatal outcome are: 0.60, 95% CI 0.51– 0.68 for abnormal UA Doppler; and 0.54, 95% CI 0.47– 0.84 for low CPR.

Conclusion

The CPR did not improve our ability to predict neonatal SGA or other short-term adverse outcomes.

Introduction

Fetal growth restriction (FGR) is a significant obstetrical complication associated with increased perinatal morbidity and mortality. Despite the implementation of fetal monitoring and surveillance by ultrasound, the assessment of fetal well-being is still challenging, even more on the fetus that is presumed to have chronic hypoxia. Currently, measurement of the umbilical artery (UA) Doppler is the standardized tool to establish the risk of adverse perinatal outcomes in the setting of fetal growth restriction, including acute hypoxia and acidosis and stillbirth¹.

Some researchers have proposed that the implementation of the Doppler measurements of the middle cerebral artery (MCA) in combination with the umbilical artery pulsatility indices, referred as the cerebroplacental ratio (CPR), may be useful in the assessment of fetal well-being for fetuses with growth restriction in order predict and avoid adverse perinatal outcomes^2–5. The mechanism of brain sparring effect will show decreased pulsatility index as part of a lower resistance system which could be associated to poor fetal prognosis^6,7. Old and recent studies suggest that the CPR improves identification of pregnancies at risk for adverse neonatal outcomes^8–10. The main aim of our study was to compare CPR to the role of UA Doppler in the prediction of neonatal SGA and short term adverse neonatal outcome in a high-risk population.

Materials and Methods

This is a multicenter prospective cohort study on women referred for fetal growth ultrasound. It was started at Washington University in St. Louis in November 2013 and completed at University of South Florida in Tampa, in July 2017. All patients provided written informed consent and were enrolled in research protocols approved by the Institutional review board at both centers.

Inclusion criteria were singleton pregnancies, gestational ages (GA) between 26.0 weeks and 36.6 weeks and dating by ultrasound prior to 24.0 weeks. Multiple gestations and fetuses with chromosomal or congenital malformations were excluded. The GA was estimated from the last menstrual period and revised by ultrasound dates when there was significant discrepancy using the American Congress of Obstetricians and Gynecologists guidelines¹¹. The biparietal diameter (BPD), head circumference (HC), abdominal circumference (AC), and femur length (FL) measurements were obtained by following the recommendations of the International Society of Ultrasound in Obstetrics and Gynecology¹².

Socio-demographic data of all enrolled subjects were collected, including maternal age, body mass index (BMI), smoking, alcohol, socioeconomic status, ethnicity, parity and coexisting maternal medical disorders, previous and current pregnancy information and ultrasound examinations.

Following delivery, we reviewed records and collected information on pregnancy complications, type of labor, mode of delivery, delivery complications, gestational weight at delivery, birth weight, Apgar scores, cord blood gases, neonatal intensive care unit (NICU) admission and discharge. Newborn outcomes were also collected including respiratory distress syndrome, intraventricular hemorrhage (IVH), hypoglycemia, necrotizing enterocolitis, periventricular leukomalacia, and seizures.

FGR was defined as an EFW <10th percentile using Hadlock standards¹³. Neonatal SGA was defined as birthweight < 10th percentile for GA by Alexander chart¹⁴.

UA and middle cerebral artery (MCA) Doppler assessments were performed. Abnormal UA Doppler was defined as: pulsatility index (PI) above the 95th percentile and absent or reverse end diastolic flow. The CPR, calculated as a ratio of the MCA PI by the UA PI, was defined as low if < 1.08^15,16. The primary outcome was the sensitivity and specificity of the two Doppler assessments to predict neonatal SGA, defined as birthweight < 10th percentile by using Alexander curves. The secondary outcomes included umbilical cord arterial pH < 7.10, Apgar scores at 5 minutes < 7, NICU admission, respiratory distress syndrome (RDS), hypoglycemia or a composite including any of these secondary outcomes.

Chi square was performed for statistical analysis using STATA 13 (College Station, Texas, USA). The risk ratios and 95% confidence intervals (CI) were estimated using the non-SGA group as the reference group. Sensitivity, specificity and positive and negative predictive values for the primary and secondary outcomes using Doppler assessments were calculated. The likelihood ratios for a positive (LR+) or negative test (LR-) were also calculated. An LR+ >10 or LR- <0.10 indicates a highly efficient predictive test.The discriminatory ability of each Doppler measurement modality was compared using area under receiver operating characteristic curves (AUC).

Results

Over the study period, 1068 singleton pregnancies were recruited, where 199 women had complete information on CPR and met the inclusion criteria. Demographic information of the current study population is detailed in Table 1. Most of the women in this study were African American (49.4%) or Caucasian (26.1%). The mean BMI was 28.5 kg/m2. Abnormal CPR was noted in 7 (3.5%) of the population. As referenced in Table 1 there were no statistically significant differences in maternal demographic characteristics between the groups with normal CPR and abnormal CPR.

Table 1.

Comparison of maternal demographic and pregnancy characteristics for cases with fetal growth restriction and cerebroplacental ratio (CPR) assessment

	Normal *CPR (n= 192)	Abnormal *CPR (n=7)	p-value
Mean age, years (±SD)	27.3 (5.7)	30.6 (4.2)	0.08
Nulliparous (%)	100 (52.1)	2 (28.6)	0.22
Ethnicity (%)			0.48
African American	80 (41.7%)	4 (57.1%)
Hispanic	23 (12%)	2(28.6%)
Asian	11 (5.7%)	0
European /white	73 (38.0%)	1 (14.2%)
Multiracial/other	5 (2.6%)	0
Mean BMI, Kg/m2 (±SD)	29.3 (2)	27.8 (6.2)	0.60
Chronic hypertension	22 (11.5%)	2 (28.6%)	0.17
Preeclampsia	19 (9.9%)	2 (28.6%)	0.11
Pre-gestational diabetes	8 (4.2%)	0	0.58
Smoking (%)	19 (9.9%)	1 (14.3%)	0.70
Gestational age at first growth ultrasound, weeks (±SD)	31.6 (2.8)	32.2 (3.1)	0.64
Gestational age at delivery, weeks (±SD)	38.0 (1.8)	34.6 (4.6)	0.09
Birthweight, grams (±SD)	2675 (518)	1890 (971)	0.07

SD= standard deviation; BMI = body mass index

UA Doppler was obtained in 208 women and 17 (8.2%) were abnormal and 191 (91.8%) were normal.

The positive (PPV) and negative (NPV) predictive values using abnormal UA Doppler to predict SGA are 70.6% and 57.1%, respectively. The PPV and NPV using low CPR to predict SGA are 57.1% and 54.4%, respectively. For the prediction of the composite adverse neonatal outcome, the PPV and NPV using UA Doppler are: 41.2% and 88.0%, respectively; and for low CPR are 42.9% and 86.9%, respectively.

Ninety-four (47.2%) had SGA and 68 (34.2%) had a composite adverse neonatal outcome. Abnormal UA Doppler showed a better sensitivity for predicting SGA (12.8%) than CPR (4.4%). The sensitivity for predicting the composite adverse neonatal outcomes using abnormal UA Doppler was also better than using CPR, 17.6% versus 7.7%, respectively (Table 2). The specificity of abnormal UA Doppler for predicting SGA was 95.6%, which is similar to that of a low CPR of 96.5% for predicting the same outcome. The specificity for predicting a composite adverse neonatal outcome was also similar. The likelihood ratio positive (LR+) for using abnormal UA Doppler to predict SGA is 2.9 and the negative LR (LR-) is 0.91. The LR+ and LR- using low CPR to predict SGA are 1.2 and 0.99, respectively. For predicting the composite adverse neonatal outcome, the LR+ and LR- using abnormal UA Doppler are 4.9 and 0.85, respectively; and using low CPR are 3.3 and 0.94, respectively.

Table 2.

Umbilical artery Doppler, low cerebroplacental ratio (CPR) and short term adverse neonatal outcomes

Abnormal UA doppler*					CPR < 1.08
	Sensitivity (%)	Specificity (%)	Relative risk	95% CI	Sensitivity (%)	Specificity (%)	Relative risk	95% CI**
SGA (n=94)	12.8	95.6	2.9	1.1–8.0	4.4	96.3	1.2	0.3–4.6
Cord pH < 7.10 (n=5)	20.0	89.8	2.0	0.3–12.2	20	96.8	6.2	0.8–45.8
5-min APGARS < 7 (n=7)	28.6	92.9	4.0	1.1–14.5	14.3	97.3	5.4	0.7–40.1
NICU admission (n=25)	28.0	95.0	5.6	2.3–13.8	13.0	97.7	5.6	1.3–23.6
RDS (n=6)	33.3	92.9	4.7	1.4–16.3	20	96.8	6.3	0.9–43.3
Hypoglycemia (n=45)	13.3	93.7	2.1	0.8–5.5	4.65	96.7	1.4	0.3–7.0
Composite (n=68)	17.6	96.4	4.9	1.8–13.5	7.7	97.7	3.4	0.8–13.8

^*PI > 95^th percentile, absent or reverse UA diastolic flow

^**95% Confidence Interval

There were 89 (44.7%) women out of 199 with EFW < 10%ile, of whom 3 (3.4%) had abnormal CPR. 10 (11%) with EFW <10% had abnormal UA. There were 3 pregnancies with EFW <10th %ile and abnormal CPR, of which 2(66.7%) had a composite adverse outcome. There were 10 pregnancies with EFW <10th %ile and abnormal UA Doppler, of which 9 (90%) had a composite adverse outcome.

The area under the ROC curve (AUC) using abnormal UA Doppler for predicting SGA was 0.54, 95% CI 0.50–0.58; and 0.51, 95% CI 0.48–0. 53 for low CPR. The AUC for predicting a composite adverse neonatal outcome are: 0.60, 95% CI 0.51–0.68 for abnormal UA Doppler; and 0.54, 95% CI 0.47–0.84 for low CPR. Figures 1 and 2 are examples of the ROC curve comparing abnormal UA Doppler to low CPR for prediction of neonatal SGA and composite adverse neonatal outcomes, respectively.

Fig 1.

ROC curve comparing abnormal UA Doppler to low CPR for predicting neonatal SGA

Fig 2.

ROC curve comparing abnormal UA Doppler to low CPR for predicting neonatal composite adverse neonatal outcomes

Discussion

Our study shows that applying the use of CPR for the prediction of adverse neonatal outcome is not superior to using umbilical artery Doppler alone. Despite the better performance of UA Doppler, the overall screening efficiency was at best modest and was poor for most outcomes evaluated.

Our findings differ from various studies suggesting that the combination of the assessments of UA and MCA Doppler has a superior role in the prediction of adverse perinatal outcomes in the setting of fetal growth restriction when compared to using these Doppler modalities separately. Similarly, other retrospective studies have shown an association between CPR and adverse intrapartum and neonatal adverse outcomes including urgent cesarean deliveries for fetal heart rate abnormalities, low cord pH, lower APGAR scores and admission to the neonatal intensive care unit^{2–4,7,17–19}. Specifically, a recent prospective study shows that the CPR < 10^th percentile has the best diagnostic performance for the prediction of low birthweight, emergency cesarean section for intrapartum fetal complications and composite adverse neonatal outcomes when used as a screening test including low risk term pregnancies prior to labor²⁰.

The strengths of this study rely on its prospective design and the use of simple, reproducible and validated ultrasound modalities. The results of the CPR test were not used for clinical management as this was not a standard of care test in both facilities, which strengthens the validity of our findings.

Some of the limitations of our study must be recognized. These include a relatively small cohort and as the prevalence of growth restriction may vary depending on the population, our findings may not be generalizable to some higher risk populations. In addition, the number of patients with abnormal CPR in our cohort was very limited. It is therefore possible that we may be underpowered to find a more positive association. Larger studies are therefore needed to address this subject before adopting the use of CPR routinely. Our findings highlight the findings from a recent meta-analysis by Oros et al, showing considerable methodological heterogeneity in studies reporting reference ranges for Doppler indices of UA, MCA and CPR. They further caution that the resulting references have important implications for clinical practice²¹. Thus, standardization of this methods is needed in order to improve their role in the management of pregnancies complicated by fetal growth restriction.

In conclusion, this small cohort study did not find the CPR to be superior to using UA Doppler for predicting adverse outcomes in pregnancies complicated by fetal growth restriction. Larger studies with well standardized definitions of abnormal CPR are needed prior to introducing CPR to routine evaluation of these pregnancies.