Volume 52, Issue 3 p. 325-331
Original Paper
Free Access

Longitudinal growth assessment for prediction of adverse perinatal outcome in fetuses suspected to be small-for-gestational age

J. Caradeux

J. Caradeux

Fetal i+D Fetal Medicine Research Center, BCNatal – Barcelona Center for Maternal-Fetal and Neonatal Medicine (Hospital Clínic and Hospital Sant Joan de Deu), IDIBAPS, University of Barcelona, Barcelona, Spain

Fetal Medicine Unit, Clínica Dávila, Santiago, Chile

Search for more papers by this author
E. Eixarch

E. Eixarch

Fetal i+D Fetal Medicine Research Center, BCNatal – Barcelona Center for Maternal-Fetal and Neonatal Medicine (Hospital Clínic and Hospital Sant Joan de Deu), IDIBAPS, University of Barcelona, Barcelona, Spain

Center for Biomedical Research on Rare Diseases (CIBER-ER), Madrid, Spain

Search for more papers by this author
E. Mazarico

E. Mazarico

Fetal i+D Fetal Medicine Research Center, BCNatal – Barcelona Center for Maternal-Fetal and Neonatal Medicine (Hospital Clínic and Hospital Sant Joan de Deu), IDIBAPS, University of Barcelona, Barcelona, Spain

Search for more papers by this author
T. R. Basuki

T. R. Basuki

Fetal i+D Fetal Medicine Research Center, BCNatal – Barcelona Center for Maternal-Fetal and Neonatal Medicine (Hospital Clínic and Hospital Sant Joan de Deu), IDIBAPS, University of Barcelona, Barcelona, Spain

Search for more papers by this author
E. Gratacos

E. Gratacos

Fetal i+D Fetal Medicine Research Center, BCNatal – Barcelona Center for Maternal-Fetal and Neonatal Medicine (Hospital Clínic and Hospital Sant Joan de Deu), IDIBAPS, University of Barcelona, Barcelona, Spain

Center for Biomedical Research on Rare Diseases (CIBER-ER), Madrid, Spain

Search for more papers by this author
F. Figueras

Corresponding Author

F. Figueras

Fetal i+D Fetal Medicine Research Center, BCNatal – Barcelona Center for Maternal-Fetal and Neonatal Medicine (Hospital Clínic and Hospital Sant Joan de Deu), IDIBAPS, University of Barcelona, Barcelona, Spain

Center for Biomedical Research on Rare Diseases (CIBER-ER), Madrid, Spain

Correspondence to: Dr F. Figueras, Maternal-Fetal Medicine Department, Hospital Clinic, University of Barcelona, Sabino de Arana 1, 08028 Barcelona, Spain (e-mail: [email protected])Search for more papers by this author
First published: 07 August 2017
Citations: 22

This article has been selected for Journal Club. Click here to view slides and discussion points.

ABSTRACT

en

Objective

Fetal growth restriction (FGR) is associated with an increased risk of adverse perinatal outcome. However, distinguishing this condition from small-for-gestational age (SGA) remains elusive. A set of criteria has been proposed recently for such a purpose, including the degree of smallness, Doppler parameters and growth velocity. The aim of this study was to establish whether the use of growth velocity adds value to Doppler assessment in predicting adverse perinatal outcome among SGA-suspected fetuses.

Methods

This was a prospective cohort study of consecutive singleton pregnancies with late (diagnosis ≥ 32.0 weeks) SGA (estimated fetal weight (EFW) < 10th centile). Longitudinal growth assessment was performed by calculation of EFW z-velocity between diagnosis and last scan before delivery. Improvement in the association with and predictive performance of EFW z-velocity for adverse perinatal outcome was compared against standard criteria of FGR evaluated before delivery (EFW < 3rd centile, abnormal uterine Doppler or abnormal cerebroplacental ratio).

Result

A total of 472 patients were evaluated prospectively for suspected SGA. Of these, 231 (48.9%) qualified as late FGR. Univariate analysis showed a significant trend towards higher frequency (14.5% vs 8.2%; P = 0.041) of EFW z-velocity in the lowest decile in pregnancies with adverse perinatal outcome. Nonetheless, the addition of EFW z-velocity improved neither the association with nor the predictive performance of standard criteria of FGR for adverse perinatal outcome.

Conclusions

Longitudinal assessment of fetal growth by means of EFW z-velocity did not have any independent predictive value for adverse perinatal outcome when used in combination with Doppler in SGA-suspected fetuses. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.

RESUMEN

es

Evaluación del crecimiento longitudinal para la predicción de un resultado perinatal adverso en fetos con sospecha de ser pequeños para la edad gestacional

Objetivo

La restricción del crecimiento fetal (RCF) está asociada con un mayor riesgo de un resultado perinatal adverso. Sin embargo, sigue siendo difícil distinguir esta condición de los fetos pequeños para la edad gestacional (PEG). Recientemente se ha propuesto un conjunto de criterios para este fin, como el grado de pequeñez, los parámetros del Doppler y la velocidad de crecimiento. El objetivo de este estudio fue establecer si el uso de la rapidez de crecimiento agrega valor a la evaluación Doppler en la predicción de resultados perinatales adversos en los fetos en los que se sospecha que son PEG.

Métodos

Este fue un estudio de cohorte prospectivo de embarazos consecutivos con feto único con PEG tardía (diagnóstico ≥32.0 semanas; peso fetal estimado (PFE) <10o percentil). Se realizó la evaluación del crecimiento longitudinal mediante el cálculo de la velocidad z del PFE entre diagnósticos y la última ecografía antes del parto. La mejora en la asociación con la velocidad z del PFE y el desempeño predictivo del resultado perinatal adverso se comparó con respecto a los criterios estándar de la RCF evaluados antes del parto (PFE <3rd percentil, Doppler uterino anómalo o relación cerebroplacentaria anómala).

Resultados

Un total de 472 pacientes fueron evaluadas prospectivamente respecto a la sospecha de ser PEG. De estas, 231 (48.9%) calificaron como RCF tardío. El análisis univariante mostró una tendencia significativa hacia una mayor frecuencia (14,5% vs 8;2%; P =0,041) de la velocidad z del PFE en el decil más bajo en embarazos con resultados perinatales adversos. No obstante, el añadir la velocidad z del PFE no mejoró ni la asociación ni el desempeño predictivo de los criterios estándar de la RCF para el resultado perinatal adverso.

Conclusiones

La evaluación longitudinal del crecimiento fetal por medio de la velocidad z del PFE no tuvo un valor predictivo independiente del resultado perinatal adverso cuando se utilizó en combinación con el Doppler en fetos con sospecha de ser PEG.

摘要

zh

通过纵向生长评估预测疑似小于胎龄儿的不良围产结局

目的

胎儿生长受限(fetal growth restriction,FGR)导致不良围产结局风险增加。然而,很难鉴别其与小于胎龄儿(small-for-gestational age,SGA)。因此,最近人们提出一系列标准,包括体重降低程度、多普勒超声参数和生长速度。本研究的目的是证实采用生长速度是否对多普勒超声评估预测疑似SGA胎儿的不良围产结局有益。

方法

本研究为前瞻性队列研究,纳入连续的迟发(≥32.0周确诊)SGA(估计胎儿体重[estimated fetal weight,EFW]<第10百分位数)单胎妊娠。通过计算确诊和分娩前最后一次超声检查间EFW z速度进行纵向生长评估。比较EFW z速度与分娩前评估的FGR公认标准相比,与不良围产结局的相关性以及预测能力的改善情况(EFW<第3百分位数,子宫多普勒超声异常或脑胎盘率异常)。

结果

共前瞻性评估472例疑似SGA的患者。其中231例(48.9%)为迟发FGR。单变量分析显示,在不良围产结局的妊娠中呈EFW z速度在最低十分位数的频率较高(14.5%和8.2%;P=0.041)的明显趋势。但是,采用EFW z速度既不能提高与不良围产结局的相关性,也不能提高FGR公认标准对不良围产结局的预测能力。

结论

在疑似SGA的胎儿中联合多普勒超声时,采用EFW z速度对胎儿生长进行纵向评估对不良围产结局无独立预测价值

INTRODUCTION

Fetal growth restriction (FGR) is an indicator of increased risk for adverse pregnancy outcome, and is associated with stillbirth, perinatal morbidity, neonatal mortality, cerebral palsy, and delayed effects into adolescence and adulthood. Suboptimal antenatal care of growth-restricted fetuses has been identified as a major cause of avoidable perinatal death1 and, accordingly, growth-restricted fetuses that are not identified prenatally show increased risks of stillbirth2 and perinatal complications3. Therefore, fetal growth assessment has become a cornerstone of routine prenatal care4.

Depending on the threshold used, 3–10% of all pregnancies have a fetal size that qualifies them as being small-for-gestational age (SGA)5. Some of these cases correspond to constitutionally, and otherwise healthy, small fetuses. Yet, a fraction of them present a pathological growth pattern, i.e. FGR, in most cases due to placental insufficiency6.

There is a growing body of evidence to suggest that, in late-onset FGR, placental insufficiency is reflected both in uterine artery Doppler7-9 and in the cerebroplacental ratio (CPR)10-12. In addition, severe smallness (usually defined as weight < 3rd centile) has also been identified as a risk factor for stillbirth13 and, independently of Doppler, for adverse perinatal outcome14.

Furthermore, a survey15 of 45 experts found fair agreement that slow growth should also be a contributory criterion to define late FGR (detected > 32 weeks). While it is conceptually sound that slow growth should qualify for FGR as it captures the dynamics of fetal growth, it is also true that few studies16-18 have compared specifically its performance against standard Doppler criteria of placental insufficiency.

The present study aimed to establish whether the use of longitudinal growth assessment by means of estimated fetal weight (EFW) z-velocity is of additional value in predicting adverse perinatal outcome among fetuses suspected to be late SGA.

METHODS

Study design and participants

Between January 2008 and December 2016, patients with a suspected SGA pregnancy at the time of referral were examined and recorded prospectively. Inclusion criteria were: normal fetal anatomy; EFW < 10th centile at inclusion, according to local reference values19; and diagnosis of SGA after 32 weeks of gestation. Exclusion criteria were: congenital malformation, chromosomopathy, infection or maternal use of an illicit substance. All pregnancies were dated according to first-trimester crown–rump length measurement20. Women at risk for abnormal fetal growth based on the identification of several maternal and pregnancy risk factors21 were offered serial ultrasound at 28, 32 and 37 weeks. Women at low risk were assessed by serial fundal-height measurement from 26 weeks and scanned at 32–34 weeks. Labor management was performed according to local protocols (Appendix S1). The local ethics committee approved the study and all women gave their informed consent to participate (IRB 2008/4422).

Measurements

In all cases, fetal biometry and prenatal fetal and maternal Doppler ultrasound examinations were performed by certified sonographers using either a Siemens Sonoline Antares (Siemens Medical Systems, Malvern, PA, USA) or a GE Voluson E8 (GE Medical Systems, Zipf, Austria) ultrasound machine equipped with a 6–2-MHz linear curved-array transducer.

The following parameters were obtained, adhering to standardized recommendations: fortnightly assessment of EFW22 derived from biparietal diameter, head circumference, abdominal circumference (AC) and femur length23; and weekly assessment of pulsatility index (PI) of the umbilical artery (UA), fetal middle cerebral artery (MCA) and uterine artery24. UA-PI and mean uterine artery PI were considered abnormal when above the 95th centile25, 26. Abnormal MCA-PI was defined as MCA-PI < 5th centile27. CPR was calculated as the ratio of MCA-PI to UA-PI28, and it was considered abnormal if < 5th centile28. EFW (the second measurement) and Doppler evaluation at the last ultrasound before delivery were considered for the analysis.

Outcomes

SGA was defined as birth weight below the 10th centile, according to customized standards29. Late FGR was defined as birth weight < 3rd centile according to local standards, or below the 10th centile plus abnormal mean uterine artery PI (> 95th centile) or abnormal CPR (< 5th centile), at the last evaluation before delivery. Pre-eclampsia was defined according to the guidelines of the International Society for the Study of Hypertension in Pregnancy30.

Adverse perinatal outcome was defined as non-reassuring fetal status requiring emergency Cesarean section, 5-min Apgar score < 7 or neonatal metabolic acidosis at birth (defined as the presence of UA pH ≤ 7.10 and base excess > 12 mEq/L at birth), need for admission to the neonatal intensive care unit or perinatal death.

Adverse neonatal outcome was defined as admission to the neonatal intensive care unit for neonatal hypoglycemia (defined as a plasma glucose level of less than 30 mg/dL in the first 24 h postpartum), neonatal hyperbilirubinemia (defined as a peak serum concentration above the 95th centile), or transient tachypnea (defined as an onset of tachypnea (> 60/min) within 6 h after birth with any of the following: expiratory grunting, flaring of the nostrils or costal retractions).

Statistical analysis

Student's t-test and Pearson's χ2 test (or Fisher's exact test) were performed for univariate comparisons between groups for quantitative and qualitative variables, respectively.

EFW measurements based on the Hadlock-422 formula were transformed to z-values (EFW z-score) according to local standards19. Longitudinal growth assessment was performed by the calculation of EFW z-velocity31 as: EFW z-score at last ultrasound within 1 week before delivery – EFW z-score at diagnosis/interval between scans (days).

The association between antenatal parameters and adverse perinatal outcome was analyzed by multivariate logistic regression. Different models were constructed for the predictors when treated as continuous and as categorical (EFW < 3rd centile, uterine artery PI > 95th centile, CPR < 5th centile, EFW z-velocity in the lowest decile) variables.

The predictive performance for adverse perinatal outcome was determined by receiver–operating characteristics (ROC) curve analysis. Paired ROC curves were compared by the DeLong method32.

All statistical analyses and graph constructions were performed using the open source software R version 2.15.1 (The R Foundation for Statistical Computing, Vienna, Austria) with the package pROC version 1.7.2. P-values < 0.05 were considered statistically significant.

RESULTS

A total of 484 patients met the inclusion criteria and were evaluated. Of these, 10 were lost to follow-up and two were excluded for premature rupture of membranes or chorioamnionitis.

A total of 472 fetuses were finally included in the analysis. At birth, 87 (18.4%) newborns had birth weight > 10th centile (but were retained in the analysis). Of the 472 suspected SGA fetuses, 231 (48.9%) met the criteria for FGR (non-exclusively): 150 (64.9%) for EFW < 3rd centile, 85 (36.8%) for abnormal uterine artery Doppler and 80 (34.6%) for abnormal CPR. The characteristics of the study population are given in Table 1.

Table 1. Maternal and pregnancy characteristics of 472 women with suspected small-for-gestational-age fetus
Characteristic Value
Age (years) 31.4 ± 5.8
BMI (kg/m2) 22.1 ± 3.2
GA at diagnosis (weeks) 34.2 ± 1.3 (32.0–38.3)
GA at last scan (weeks) 38.0 ± 1.3 (34.0–42.1)
GA at delivery (weeks) 38.9 ± 1.2 (34.1–42.1)
Mean interval between US examinations (days) 26.6 ± 10.9 (14–56)
White European ethnicity 324 (68.6)
Low socioeconomic level 118 (25.0)
Nulliparous 292 (61.9)
Smoker 104 (22)
Previous adverse perinatal outcome 14 (2.9)
Maternal disease* 24 (5.1)
Pre-eclampsia 41 (8.7)
Labor induction 296 (62.7)
Cesarean section 137 (29)
For fetal distress 45 (32.8)
Elective 53 (38.6)
For induction failure 39 (28.5)
Adverse perinatal outcome 131 (27.8)
Emergency Cesarean section for NRFS 45 (9.5)
5-min Apgar score < 7 3 (0.6)
Neonatal acidosis 15 (3.2)
NICU admission 83 (17.6)
Perinatal death 0 (0)
  • Data are presented as mean ± SD, mean ± SD (range) or n (%).
  • * Chronic hypertension, diabetes mellitus, renal disease, autoimmune disease or coagulation disorder.
  • Arterial pH < 7.10 and base excess > 12 mEq/L.
  • BMI, body mass index; GA, gestational age; NICU, neonatal intensive care unit; NRFS, non-reassuring fetal status; US, ultrasound.

Table 2 summarizes the ultrasound and Doppler findings at the last ultrasound assessment, according to the occurrence of adverse perinatal outcome. As expected, those with adverse perinatal outcome delivered earlier, had lower EFW, higher mean uterine artery PI and UA-PI values, and had lower MCA-PI and CPR values. Furthermore, those with adverse perinatal outcome had a higher proportion of abnormal Doppler assessment for all vessels evaluated. Of note, EFW z-velocity did not differ between cases with and those without adverse perinatal outcome (−0.01 ± 0.19 vs −0.04 ± 0.20 (mean ± SD); P = 0.236) (Figure 1). There was a significant trend towards higher frequency of EFW z-velocity in the lowest decile in pregnancies with adverse perinatal outcome compared with those without (14.5% vs 8.2%; P = 0.041).

Table 2. Assessment of pregnancies with suspected small-for-gestational-age (SGA) fetus, according to presence or absence of adverse perinatal outcome
Adverse perinatal outcome
Assessment No (n = 341) Yes (n = 131) P
GA at diagnosis (weeks) 34.3 ± 1.4 34.2 ± 1.2 0.524
GA at delivery (weeks) 39.0 ± 1.2 38.5 ± 1.3 < 0.001
EFW at diagnosis (g) 1866 ± 245 1828 ± 234 0.123
EFW z-score at diagnosis −1.8 ± 0.4 −1.9 ± 0.5 0.023
EFW at last US examination (g) 2434 ± 290 2345 ± 315 0.004
EFW z-score at last US examination −1.9 ± 0.6 −2.0 ± 0.7 0.044
Mean interval between US examinations (days) 26.9 ± 11.2 25.8 ± 10.1 0.36
Mean uterine artery PI z-score 0.2 ± 1.4 0.6 ± 1.5 0.007
Mean uterine artery PI > 95th percentile 48 (14.1) 37 (28.2) < 0.001
UA-PI z-score 0.8 ± 0.4 1.0 ± 0.5 0.003
UA-PI > 95th percentile 1 (0.3) 5 (3.8) 0.007
MCA-PI z-score 0.1 ± 1.1 −0.2 ± 1.1 0.004
MCA-PI < 5th percentile 13 (3.8) 12 (9.2) 0.02
CPR z-score −0.6 ± 1.1 −0.9 ± 1.1 0.001
CPR < 5th percentile 45 (13.2) 35 (26.7) < 0.001
EFW z-velocity −0.01 ± 0.19 −0.04 ± 0.2 0.236
EFW z-velocity in lowest decile 28 (8.2) 19 (14.5) 0.041
SGA* 260 (76.2) 125 (95.4) < 0.001
FGR 151 (44.3) 80 (61.1) 0.001
  • Data are given as mean ± SD or n (%).
  • * Birth weight < 10th centile.
  • Birth weight < 3rd centile, or < 10th centile with abnormal uterine artery pulsatility index (PI) or abnormal cerebroplacental ratio (CPR), at last evaluation before delivery.
  • EFW, estimated fetal weight; FGR, fetal growth restriction; GA, gestational age; MCA, middle cerebral artery; UA, umbilical artery; US, ultrasound.
Details are in the caption following the image
Estimated fetal weight z-velocity in pregnancies with suspected small-for-gestational-age fetus, in those without (a) and those with (b) adverse perinatal outcome.

Table 3 and Figure 2 show the multivariate analysis of the association of antenatal assessment parameters with adverse perinatal outcome. Either as continuous or categorical predictive variables, only Doppler parameters were associated significantly and independently with adverse perinatal outcome.

Table 3. Logistic regression analysis for prediction of adverse perinatal outcome in pregnancies with suspected small-for-gestational-age fetus, for standard approach alone and that incorporating estimated fetal weight (EFW) z-velocity, according to variable type
Variable type, approach and parameter OR (95% CI) P
Continuous
Standard
EFW z-score at last US 0.81 (0.58–1.12) 0.204
Mean uterine artery PI 1.16 (1.01–1.33) 0.032
CPR 0.74 (0.60–0.90) 0.003
Standard + EFW z-velocity
EFW z-score at last US 0.79 (0.52–1.21) 0.283
Mean uterine artery PI 1.16 (1.01–1.33) 0.033
CPR 0.74 (0.60–0.90) 0.003
EFW z-velocity 1.10 (0.29–4.21) 0.888
Dichotomous
Standard
EFW < 3rd percentile at last US 1.49 (0.96–2.30) 0.074
Mean uterine artery PI > 95th percentile 2.04 (1.23–3.38) 0.006
CPR < 5th percentile 2.04 (1.22–3.41) 0.007
Standard + EFW z-velocity
EFW < 3rd percentile at last US 1.36 (0.85–2.19) 0.204
Mean uterine artery PI > 95th percentile 2.05 (1.24–3.40) 0.005
CPR < 5th percentile 2.00 (1.19–3.35) 0.009
EFW z-velocity in lowest decile 1.39 (0.69–2.81) 0.354
  • CPR, cerebroplacental ratio; OR, odds ratio; PI, pulsatility index; US, ultrasound.
Details are in the caption following the image
Multivariate analysis for association of antenatal assessment parameters (as categorical predictive variables) with adverse perinatal outcome in pregnancies with suspected small-for-gestational-age fetus. CPR, cerebroplacental ratio; EFW, estimated fetal weight; OR, odds ratio; p3, 3rd centile; p5, 5th centile; p95, 95th centile; PI, pulsatility index.

Regarding the prediction of adverse perinatal outcome, when performed using predictors as continuous variables, areas under the curve (AUC) of the standard approach (EFW z-score, mean uterine artery PI and CPR) alone and in combination with EFW z-velocity were 0.622 (95% CI, 0.564–0.68) and 0.623 (95% CI, 0.565–0.68), respectively (P = 0.35). When used as categorical variables, AUCs of the standard approach alone and in combination with EFW z-velocity were 0.614 (95% CI, 0.558–0.669) and 0.617 (95% CI, 0.561–0.672), respectively (P = 0.50).

Additionally, logistic regression analyses were performed to evaluate if longitudinal assessment of fetal growth by means of z-velocity was associated with a higher risk of adverse neonatal outcome (Table S1). Either as a continuous or as a categorical predictive variable, EFW z-velocity did not have any independent association with adverse neonatal outcome when used in combination with Doppler for SGA-suspected fetuses.

DISCUSSION

Slow growth has been proposed as a contributory criterion to define late FGR15. However, its value in the prediction of adverse perinatal outcome over other standard criteria of late FGR33 has been evaluated scarcely. In our study on suspected late-onset SGA, longitudinal assessment of fetal growth by means of z-velocity did not add value to Doppler criteria (CPR and uterine artery Doppler) in predicting adverse outcome.

Based on recent evidence34, 35, many guidelines21, 36, 37 recommend delivering all suspected SGA babies at 37 weeks, even though, under this definition, many constitutionally small babies are included. By curtailing pregnancy by 3 weeks, a few instances of late stillbirth could be prevented, but this should be weighed against the risks of early term birth38, 39. Thus, identifying small fetuses at risk of adverse outcome is key for stratification and tailored management, which has the potential to avoid many unnecessary interventions.

Some previous reports in high-risk populations40-42 have shown that defective longitudinal fetal growth is associated with adverse perinatal outcome. In these series, ‘high-risk’ was defined variably by maternal characteristics, obstetric history or EFW < 10th centile. More recently, two studies used serial ultrasound evaluations of fetal growth as a strategy to improve diagnosis of FGR. The first, conducted by Sovio et al.17 in a large cohort of 3977 unselected nulliparous women, found a significant association of second- to third-trimester slow growth velocity with adverse outcome. On stratification, the association remained significant only for those cases (n = 560) with EFW < 10th centile (relative risk, 1.96; 95% CI, 1.21–3.19). The effect of SGA on adverse outcome was modified significantly by the presence of low growth velocity, to a greater extent than by the presence of abnormal UA or uterine artery Doppler. CPR was not available for analysis. On the other hand, Karlsen et al.18 in a prospective cohort of 211 pregnancies at risk of placental dysfunction (24.6% of them with EFW < 5th centile) aimed at evaluating whether the use of conditional growth centiles could improve the prediction of adverse outcome. They found that size centiles and conditional growth centiles contribute independently to the prediction of adverse outcome and their combination further improves the prediction model. The combination of growth velocity with Doppler parameters for predicting adverse outcome was not explored.

Recently, Gordijn et al.15 made a significant contribution to reaching a consensus-based definition of FGR, including biometric (both cross-sectional and longitudinal) as well as Doppler parameters. Yet, several aspects of longitudinal assessment are still to be elucidated. First, the proposed threshold for qualifying slow growth (AC/EFW crossing two quartiles) is arbitrary and lacking an interval time frame; the same amount of decline over a short time interval would be seen as more concerning. There are, however, concerns that intervals between examinations of less than 3 weeks may result in an exceedingly high number of false positives43. Another issue is whether the baseline measurement from which the growth rate is to be evaluated should be set in the second trimester (when the growth potential could be assumed to be still intact44) or in the third trimester (when the measurement error is expected to be lower45). Second, drawing conclusions on longitudinal growth by plotting time-point measurements of size is also arguable46, even when appropriate charts are used. To overcome this limitation, several methods have been proposed (conditional growth47, growth rate41 or velocity16, or individualized growth trajectories48) without clear evidence of which of them is superior. Third, rather than evaluating the performance of growth velocity as a standalone predictor, its predictive capacity over other biometric and functional parameters should be considered. In our study, although there was a significant trend towards higher frequency of EFW z-velocity in the lowest decile in pregnancies destined to have adverse perinatal outcome, when multivariate analysis was performed, growth velocity did not show independent association or predictive performance. It could be speculated that placental insufficiency is already captured by abnormal Doppler (either in the maternal or in the fetal compartment, i.e. uterine Doppler and CPR, respectively) rendering longitudinal growth devoid of independent predictive capacity for adverse outcome. Indeed, recent evidence49, 50 on retrospective series shows that there is a correlation between growth rate and CPR. However, on assessment of the association with adverse outcome, only the Doppler parameter remained independently associated49.

In the Delphi study15, uterine artery Doppler was not selected as a criterion of fetal growth restriction. However, some studies7, 9, 51 suggest that uterine artery Doppler performed at the time of diagnosis of SGA is an independent predictor of adverse outcome, irrespective of other markers. The findings of the current study are in keeping with this evidence.

To the best of our knowledge, our series of suspected SGA fetuses with longitudinal growth assessment and full Doppler evaluation is the largest published so far. However, we acknowledge some limitations of our study. First, our findings could not be translated to FGR occurring early in pregnancy, as we only included late-onset cases (as defined in the recent Delphi consensus15 as diagnosis after 32 weeks). However, in early-onset FGR, the main clinical challenge is management rather than detection. Second, only two measurement points (the first and last ultrasound) were used to derive growth rate; additional measurements could improve detection, although the evidence for this is lacking. Third, among the available longitudinal methods, we used only z-velocity, as it is more intuitive and easy to handle in clinical practice. Fourth, we were not able to asses if normal growth velocity was influenced by maternal characteristics; therefore, we cannot rule out that expected growth velocity varied between patients. Fifth, it could be argued that the use of AC z-velocity instead of EFW z-velocity could have yielded different results. However, there is conflicting evidence regarding this issue and, in clinical practice, EFW is evaluated consistently before and after delivery, since weight and not AC is the neonatal standard for growth. Finally, we reckon that, due to our definition of late FGR (necessarily including a birth weight below the 10th centile), some cases with normal size but Doppler signs of placental insufficiency may have been missed.

In conclusion, longitudinal assessment of fetal growth by means of EFW z-velocity does not add value to Doppler criteria in predicting adverse perinatal outcome in fetuses suspected to be SGA.

Acknowledgments

This project has been funded with support of the Erasmus+ Program of the European Union (Framework Agreement number: 2013-0040). This publication reflects the views only of the authors, and the Commission cannot be held responsible for any use that may be made of the information contained herein. Additionally, the research leading to these results has received funding from ‘la Caixa’ Foundation; Cerebra Foundation for the Brain Injured Child (Carmarthen, Wales, UK) and AGAUR 2014 SGR grant no. 928.