Reduced growth velocity from the mid-trimester is associated with placental insufficiency in fetuses born at a normal birthweight

This study aimed to establish whether growth velocities in AGA fetuses best reflect placental function if measured between 20–28, 28–36, and 20–36 weeks’ gestation. We analysed data from participants of the Fetal Longitudinal Assessment of Growth (FLAG) study—a prospective cohort study conducted at a tertiary maternity hospital in Melbourne, Australia, which has been previously described in detail [11]. In brief, the FLAG study prospectively recruited nulliparous women who underwent blinded ultrasound scans at 28 and 36 weeks’ gestation. From these scans, 28–36-week estimated fetal weight (EFW) and abdominal circumference (AC) growth velocities were calculated, and we have reported that growth velocity correlates with antenatal, intrapartum and postnatal clinical indicators of placental insufficiency among infants born AGA [11]. The FLAG study was approved by the Mercy Health Research Ethics Committee (Ethics Approval Number R14/12) and written informed consent was obtained from all participants.

For this study, second trimester fetal size was determined using biometry measurements reported at the participants’ routine mid-trimester morphology ultrasound scan. One hundred eighty-seven (53.9%) of these were performed at the Mercy Hospital for Women, with the remainder being performed by external medical imaging providers. Gestation-dependent centiles were assigned for both mid-trimester EFW and AC. EFW and AC centile changes between 20–28 and 20–36 weeks were calculated and examined for associations with the same antenatal, intrapartum and postnatal indicators of placental insufficiency. These correlations were then compared to those previously reported for fetal growth velocities between 28 and 36 weeks [11].

This study was designed to investigate whether slowed growth velocities from the mid-trimester are associated with features of placental insufficiency among AGA infants delivered at term. Therefore, SGA infants (customised birthweight < 10th centile) were excluded from  the main analyses.

Three hundred forty-seven of the participants recruited to the FLAG study completed both 28- and 36-week research ultrasound examinations and were therefore eligible for inclusion. Biometric measurements (biparietal diameter, head circumference, AC and femur length) were obtained from the mid-trimester morphology ultrasound and the EFW was derived from the Hadlock equation using all four parameters or from the Shepard formula utilising three parameters for two cases where there were missing femur length measurements [13, 14]. The mid-trimester scans had been performed between 17⁺⁶ and 23⁺⁰ weeks’ gestation. Twenty-eight- and 36-week fetal biometry measurements were obtained as previously described [11].

As in the original FLAG study, gestation-specific mid-trimester ultrasound EFW centiles were customised using the GROW software (http://​www.​gestation.​net/​) [15]. The GROW software was used to adjust for maternal height, weight and parity, fetal sex and gestational age. We did not customise for maternal ethnicity as it is known to be inconsistently assigned among our patient population [16]. The GROW software only assigns customised centiles from a minimum of 20 weeks’ gestation. For the 39% of participants who had their mid-trimester scan prior to 20 weeks, we calculated the Hadlock EFW gestation-specific centile at the gestation of their mid-trimester scan, and then calculated the EFW at 20 weeks’ gestation that would correlate to that same Hadlock centile. We then used the extrapolated 20-week EFW to calculate the customised centile using the GROW software. Coefficients for the Australian application of GROW were informed by a local dataset. All AC centiles were assigned using the Chitty AC equation [17].

We calculated the change in EFW centile between 20 and 28 weeks, by subtracting the 20-week customised EFW centile from the 28-week customised EFW centile. Similarly, we calculated the change in EFW centile between 20 and 36 weeks by subtracting the 20-week EFW centile from the 36-week centile. The same process was used to calculate AC growth velocities for both time epochs. A fetus with a reduction in EFW or AC centile over the time period thus had a negative number to describe the growth velocity, a fetus with no change in centile had a velocity value of zero and a fetus with an increase in centile across gestation had a positive velocity value.

To standardise comparison of growth velocities across the cohort, the change in both EFW and AC centiles between ultrasounds were divided by the exact number of days between examinations, to create a centile change per day. For the 20–28-week velocities, we then multiplied this number by 56 to facilitate standardised comparison of centile change over exactly 8 weeks. For the 20–36-week velocities, we multiplied this number by 112 to compare change over exactly 16 weeks. Consistent with our previous study [11], we defined low growth velocity as a fall in EFW or AC of at least 30 centiles between the two ultrasound scans to allow for the maximal ultrasound error (approximately 15%) when estimating fetal weight.

The outcomes used as antenatal, intrapartum and postnatal evidence of placental insufficiency among AGA fetuses were as follows: cerebroplacental ratio (CPR) < 5th centile at 36 weeks—reflecting increasing placental resistance and redistribution of blood flow to the brain—a known fetal adaptation to relative hypoxia [18], neonatal acidosis (umbilical artery pH < 7.15 at birth) following the hypoxic challenge of labour—reflecting reduced placental reserve [19] and low neonatal body fat percentage (BF%) measured by air displacement plethysmography (ADP)—reflecting reduced in utero nutritional reserve (BF% < 4.2% for males and < 5.8% for females) [20]. Where possible, study infants had ADP performed within the first 4 days after birth prior to their discharge from hospital. ADP assessments in our cohort were performed at a mean (standard deviation) 44.7 (12.1) hours of age. Untrimmed placentas were weighed in the birth suite and were classified as either < 10th centile or ≥ 10th centile adjusted for gestation at birth and infant sex [21]. A placenta < 10th centile for gestation and infant sex was considered a further indicator of placental insufficiency as smaller placentas are associated with FGR [22].

We also examined whether EFW and AC growth velocities between 20 and 28 weeks were associated with clinical indicators of placental insufficiency. When linear regression was performed, 20–28-week growth velocities were not significantly correlated with CPR MoM (Fig. 3a, b). EFW and AC 20–28-week growth velocities positively correlated with all neonatal body fat measures—Ponderal Index, neonatal BF% estimated by skinfold testing and ADP BF%, except 20–28-week AC growth velocity and ADP BF% (Fig. 3c–h), but with weaker correlations than those demonstrated by 20–36-week growth velocities. Using logistic regression, we did not identify significant relationships between 20- and 28-week EFW and AC growth velocities and any of the antenatal, intrapartum or neonatal indicators of placental insufficiency (Table 1) except for 20–28 week AC growth velocity and intrapartum acidosis. For every one centile decrease in AC velocity, the odds of intrapartum acidosis increased by 2.1% (95% CI 0.2–4.1%, P = 0.03). However, when we dichotomised the cohort, low growth velocity defined as 20–28 week AC velocity of < − 30 centiles was not associated with significantly increased risk of intrapartum acidosis (RR (95% CI) = 1.5 (0.4–4.5), P = 0.64).

Bild vergrößern

We report that reduced EFW and AC growth velocities across the gestational epoch of 20–36 weeks are associated with antenatal, intrapartum and neonatal indicators of placental insufficiency in infants born AGA. We found significant correlations between reduced 20–36-week growth velocities and (i) low CPR at 36 weeks’ gestation, reflective of adaptive fetal cerebral redistribution and increased placental resistance; (ii) development of acidosis under the hypoxic challenge of labour; (iii) low neonatal body fat stores; and (iv) small placentas with weight < 10th centile for infant sex and gestation, even though the majority of fetuses with low growth velocity did not exhibit these features of placental insufficiency. This supports our previous findings suggesting that AGA fetuses with reduced growth velocity across the third trimester may have experienced placental insufficiency, placing them at increased risk of stillbirth [11, 24] despite their ‘normal’ birthweight. This study reports that the routine mid-trimester ultrasound can be used as the baseline assessment of fetal size from which a reliable measure of slowing growth can be generated at 36 weeks’ gestation. Slowing of growth across this gestational epoch is associated with features of placental insufficiency at term.

Importantly, these findings suggest this novel assessment of fetal wellbeing could be feasibly achieved with the addition of just a single 36-week scan to standard antenatal care. We report that fetal growth velocity between 20 and 36 weeks compares favourably to our previously reported associations of 28–36-week fetal growth velocity and features of placental insufficiency [11], potentially obviating the need for an additional 28-week ultrasound. Furthermore, this adds to the previously established benefits of a routine 36-week scan: increased detection of SGA, large-for-gestational-age and breech presentation at term [12, 25‐28]. Although a large randomised controlled trial of routine third trimester ultrasound did not demonstrate a reduction in severe adverse perinatal outcomes [28], the additional information provided from fetal growth velocity assessment may add value to the 36-week scan. Identifying AGA fetuses with slowed growth velocity may unmask a currently unrecognised group experiencing placental insufficiency, who are potentially at increased risk of stillbirth.

We did not demonstrate strong relationships between 20- and 28-week growth velocities and indicators of placental insufficiency at term. This finding is in keeping with the hypothesised pathophysiology of late-onset FGR which typically occurs at or beyond 32 weeks’ gestation [9, 29]. Placental insufficiency occurring early enough in pregnancy to cause growth decrements during the mid-trimester would likely culminate in the birth of an infant preterm, SGA or both. Such infants were excluded from our analyses given they are already known to represent high-risk cohorts.

This study supports the findings of the Pregnancy Outcome Prediction study, where low 20–36-week AC growth velocity was associated with poorer perinatal outcomes among fetuses with an EFW < 10th centile at 36 weeks [12]. In the Pregnancy Outcome Prediction study, SGA fetuses with AC growth velocity in the lowest decile had a RR of 2.5 for neonatal morbidity, compared to a RR of only 1.3 for SGA fetuses who had exhibited normal AC growth velocity from the mid-trimester. Similar associations were observed for other adverse outcomes, such as metabolic acidosis, neonatal unit admission and severe adverse perinatal outcome. This highlights the added value of fetal growth velocity assessment to size among SGA fetuses in late pregnancy.

Our findings suggest that 20–36-week growth velocity is also an important indicator of wellbeing among AGA fetuses. We have previously reported significant associations between 28- and 36-week growth velocity and antenatal, intrapartum and postnatal indicators of placental insufficiency [11] in a prospective cohort of low risk nulliparous women who delivered AGA infants. The current study demonstrates that these growth trajectory assessments could start from the time of routine mid-trimester ultrasound examination, given that 20–36-week fetal growth velocities predict indicators of placental insufficiency as well as those calculated between 28 and 36 weeks. A previous study of 934 AGA infants also reported significant inverse associations between abdominal circumference velocity from the mid-trimester and (i) composite adverse neonatal outcome and (ii) Neonatal Intensive Care admission [30]. An important limitation however was its retrospective design. The third trimester data came from clinically indicated referral ultrasounds, where the biometry was not blinded, and the results were used to inform clinical care. In contrast, our prospective study suggests that these findings may be generalisable to unselected populations.

Major strengths of this study include its prospective design and the examination of multiple, robust measures of placental insufficiency across the antenatal, intrapartum and neonatal periods. CPR, umbilical artery pH, ADP BF% and small placental size are each reliable indicators of placental insufficiency [19, 22, 31‐33], and the relationships between growth velocity and these outcome measures were consistent. The study was also strengthened by our use of customised EFW and birthweight centiles, which have a stronger association with adverse perinatal outcomes than standard population references [34]. By calculating centile change per day, we were able to perform standardised comparisons. By then dichotomising the cohort according to a definition of low growth velocity (loss of 30 centiles between 20 and 36 weeks), we have aided the clinical interpretation, and potential application, of our data.

A limitation of the study is that we were unable to assess for intra-observer variability among fetal biometry measurements recorded at the mid-trimester morphology scans. Additionally, there was greater variability in the timing and locations of the mid-trimester ultrasound scans (17⁺⁶ to 23⁺⁰ weeks inclusive) when compared to the standardised 28- and 36-week research scans (27⁺⁰ to 29⁺⁰ and 35⁺⁰ to 37⁺⁰ respectively). Despite the greater heterogeneity of the mid-trimester scans, the relationships observed between growth velocities measured from 20 to 36 weeks, versus 28–36 weeks, and indicators of placental insufficiency were remarkably similar (Table 1). That the mid-trimester morphology scans were performed at more variable sites and gestations reflects real-life clinical practice and potentially increases the generalisability of our findings. In addition, there were no significant differences in the median gestations of the mid-trimester scans, nor the AC or EFW centiles derived from these depending on whether they were performed internally or by external providers (Additional file 1: Table S1). That we were unable to perform histopathology on the placentas of study participants is a further limitation. Finally, the study was limited by its cohort size. It was underpowered to detect important but uncommon perinatal outcomes such as stillbirth. Results from this study warrant validation in a larger study.

This study suggests that calculation of fetal growth velocity between the routine mid-trimester ultrasound and 36 weeks could help identify undetected placental insufficiency among AGA fetuses. It is plausible that these fetuses may represent an unsuspected cohort at increased risk of stillbirth. Fetuses with low growth velocity may benefit from increased surveillance and timely delivery to reduce the risks associated with placental insufficiency, including stillbirth. This study suggests that—with the addition of only one 36-week scan to routine antenatal care—identification of FGR among the AGA could be increased. Given that universal late-pregnancy ultrasound has already been established to (i) almost triple the identification of SGA, (ii) identify which SGA fetuses are at greatest risk of adverse outcome, (iii) almost eliminate undiagnosed breech presentation and its associated mortality and (iv) be cost effective [12, 26, 35], this additional assessment could be feasibly incorporated into clinical care. Secondly, the increases in risk reported here associated with loss of greater than 30 or more EFW or AC centiles may assist clinicians interpreting fetal growth velocity when caring for pregnant women who have undergone serial ultrasounds scans where the fetus is not—or not yet—SGA. Here, we have demonstrated the feasibility of generating growth velocity per day values. From these, a fixed multiplier can be used to standardise comparison between scans, irrespective of the exact gestations at which they are performed. Finally, by identifying that the baseline assessment to calculate fetal growth velocity can be undertaken as early as the mid-trimester, this study proposes a novel application of the routine morphology ultrasound.

These findings suggest that a fall of 30 or more EFW or AC centiles between the 20- and 36-week scans could be used as a pragmatic clinical cut-off to define low fetal growth velocity, given the consistent associations with markers of placental insufficiency. Future research will involve testing this threshold in a larger prospective study to more comprehensively determine its diagnostic performance for adverse perinatal outcomes. If confirmed, AGA fetuses with reduced growth velocity may be candidates for increased surveillance and timely delivery to reduce stillbirth risk.