Abdominal circumference growth velocity as a predictor of adverse perinatal outcomes in small-for-gestational-age fetuses

Abstract

Objective

To assess the predictive value of abdominal circumference growth velocity (ACGV) between the second and third trimesters to predict adverse perinatal outcomes in a cohort of small-for-gestational-age fetuses without evidence of placental insufficiency (i.e. fetal growth restriction).

Material and methods

This is a single-center retrospective cohort study of all singleton pregnancies with small-for-gestational-age fetuses diagnosed and delivered at a quaternary institution. Crude and adjusted odds ratios (ORs) and corresponding confidence intervals (CIs) were calculated via logistic regression models to assess the potential association between abnormal ACGV (i.e. ≤10th centile) and adverse perinatal outcomes defined as a composite outcome (i.e. umbilical artery pH <7.1, 5-min Apgar score <7, admission to the neonatal intensive care unit, hypoglycemia, intrapartum fetal distress requiring expedited delivery, and perinatal death). Furthermore, the area under the receiver-operating characteristic curve (AUC) of three logistic regression models based on estimated fetal weight and ACGV for predicting the composite outcome is also reported.

Results

A total of 154 pregnancies were included for analysis. The median birthweight for the cohort was 2,437 g (interquartile range [IQR] 2280, 2635). Overall, the primary composite outcome was relatively common (29.2%). In addition, there was a significant association between abnormal ACGV and adverse perinatal outcomes (OR 3.37, 95% CI 1.60, 7.13; adjusted OR 4.30, 95% CI 1.77, 10.49). Likewise, the AUC for the ACGV was marginally higher (0.64) than the estimated fetal weight (0.54) and ACGV + estimated fetal weight (0.54). Still, no significant difference was detected between the curves (p = 0.297).

Conclusions

Our results suggest that an ACGV below the 10th centile is a risk factor for adverse perinatal outcomes among small-for-gestational-age fetuses.

Keywords:

• Fetal growth
• growth velocity
• abdominal circumference growth velocity
• small-for-gestational-age
• perinatal outcomes
• estimated fetal weight

Introduction

Low birth weight represents one of the leading causes of adverse perinatal outcomes (APO) worldwide and has been recently considered a risk factor for disease in adult life [1–3]. Conventionally, a fetus is defined as small-for-gestational-age (SGA) when the estimated fetal weight (EFW) is below the 10th centile for gestational age [4]. However, this group comprises a heterogeneous population. Growth-restricted fetuses (i.e. those with a pathological process and who have failed to reach their genetic growth potential) should be differentiated from those constitutionally small fetuses [5]. The former group is at increased risk of perinatal mortality and adverse neurodevelopmental outcomes; the latter, while small, are otherwise considered healthy [6]. This distinction reduces adverse outcomes in growth-restricted fetuses and prevents unnecessary interventions in constitutionally small fetuses [7,8].

Identification of growth-restricted fetuses is only sometimes straightforward. Several tools have been used to help differentiate those fetuses failing to reach their growth potential from those constitutionally small (e.g. Doppler evaluation of placental and fetal circulation, customized growth charts, biomarkers). Nevertheless, none of these tools has proved sufficient as a standalone criterion for the differential diagnosis [8]. Longitudinal fetal growth assessment has recently been considered a tool for diagnosing, surveillance, and managing fetal growth anomalies [6]. As opposed to size, an individual set of anthropometric measurements assessed at a specific time, growth is a change in a measure per unit of time, hence a dynamic process. Assessing a growth trajectory could aid in identifying those fetuses failing to reach their growth potential [9]. Abdominal circumference growth velocity (ACGV), a proxy of longitudinal fetal growth, has been studied as a predictor of APO in different populations [10–12]. The Pregnancy Outcome Prediction (POP) study, a prospective and blinded study in unselected nulliparous women, found that ACGV predicts poor outcomes in SGA fetuses [7]. However, it should be noted that most studies of ACGV in SGA populations include fetuses with EFW below the third centile or clear evidence of placental insufficiency (i.e. abnormal Doppler evaluation), which should be considered growth-restricted fetuses. Therefore, the clinical value of ACGV in identifying pathological growth in the group of SGA fetuses without evidence of placental insufficiency (i.e. EFW between the 3^rd and 10th centiles with normal Doppler evaluation) remains unclear. The present study aimed to assess the presence and potential association of ACGV between the second and third trimesters with APO in SGA fetuses without evidence of placental insufficiency.

Material and methods

This single-center retrospective cohort study was conducted at the Maternal-Fetal Medicine Department of the National Perinatology Institute in Mexico City from January 2016 through August 2022. Women with a singleton pregnancy and a diagnosis of SGA fetus (i.e. EFW between the 3–10th centile for gestational age and normal Doppler evaluation [i.e. umbilical artery pulsatility index <95th centile, middle cerebral artery pulsatility index >5th centile, cerebroplacental ratio >5th centile, mean uterine artery pulsatility index <95th centile]) at third-trimester scan (35–36.5 weeks’ gestation) were included [8,13–15]. A reliable last menstrual period for gestational age was confirmed based on a first trimester (i.e. cephalocaudal length) or an early second-trimester scan (i.e. biparietal diameter). Fetuses with lethal structural anomalies or requiring neonatal surgery, chromosomal abnormalities, and those with incomplete medical records were excluded.

All scans were performed by maternal-fetal medicine specialists on commercially available equipment (General Electric Voluson 730 Expert, Voluson E8) and according to standardized guidelines for fetal biometry [16]. EFW was calculated from the equation proposed by the INTERGROWTH-21st project [17,18]. ACGV was defined as the Z score (i.e. the number of standard deviations [SD] that a measurement is from the mean) difference of the abdominal circumference (AC) between the second (19.0–21.6 weeks of gestation) and third (35–36.5 weeks of gestation) trimesters, divided by the time interval between the two measures expressed in days, and multiplied by 100. The ACGV centile was estimated according to the reference ranges published by Vannuccini et al. using an online calculator (www.calculosaurus.com) [19]. An ACGV equal to or lower than the 10th centile was considered abnormal [20]. The monitoring, time, and mode of delivery of all pregnancies were at the discretion of the attending obstetrician.

The primary outcome was a composite APO, defined as the presence of one or more of the following pH <7.1, Apgar at 5 min <7, neonatal intensive care unit admission in the first 48 h of life, hypoglycemia (<45 mg/dl [2. 5 mmol/l] in the first 24 h of life), intrapartum fetal distress (defined as category II cardiotocographic monitoring trace not recovering after intrauterine resuscitation maneuvers or category III) requiring expedited delivery, and perinatal death. Demographic data, medical history, fetal biometric parameters, and perinatal outcomes were collected from the medical records.

Statistical analysis

Logistic regression models were used to calculate crude and adjusted odds ratios (ORs) and their corresponding confidence intervals (CIs) to assess the presence and magnitude of a potential association between abnormal ACGV and APO. The following covariates were included for the adjusted effects analyses, maternal age, pregestational body mass index, preeclampsia, SGA or fetal demise history, tobacco exposure, conception method, hypertension, diabetes, lupus, antiphospholipid syndrome, thrombophilia, and hypothyroidism. The areas under the receiver-operating-characteristic curve (AUC) of three logistic regression models based on EFW and ACGV for the prediction of APO are also reported. Furthermore, areas under ROC curves were compared using the method proposed by DeLong et al. [21]. Potential differences in the study population were appraised using standardized differences. A significant P-value was defined as <0.05. As an exploratory analysis, we also evaluated the potential association of an ACGV equal or lower to the 20th centile with APO. Data are presented as mean (standard deviation [SD]), median (interquartile range [IQR]), or n (%). The statistical analysis was performed using Stata version 15.1 (StataCorp LLC, Texas, USA).

Ethics statement

This was a retrospective analysis of anonymized data, so no formal ethics committee review and approval were necessary according to our institutional regulations. Of note, all women give their written consent to use their routinely collected hospital data for retrospective studies without patient identifiers.

Results

During the study period, 154 women met the inclusion criteria and were included for analysis. The mean gestational age at the second and third-trimester scans was 20.64 ± 0.88 and 35.99 ± 0.55, respectively. Women were divided into two groups according to normal or abnormal ACGV; 49.4% (76) had an ACGV >10th centile, while 50.6% (78) had an abnormal ACGV (≤10th centile). Maternal demographic and obstetric characteristics are presented in Table 1; there was a low frequency of maternal comorbidities in the entire cohort.

Table 1. Maternal demographic characteristics, obstetric history, and comorbidities.

Characteristic	Total N = 154	ACGV^a ≤10^th centile N = 78	ACGV >10^th centile N = 76	Standardized difference
Maternal age, years, mean, SD^b	28.92 ± 7.64	29.56 ± 7.37	28.25 ± 7.90	−0.172
Pregestational BMI^c, kg/m^2, mean, SD	25.43 ± 4.56	25.86 ± 4.72	24.98 ± 4.38	−0.193
Pre-eclampsia history, n (%)	13 (8.4)	7 (9)	6 (7.9)	0.038
Low birthweight history, n (%)	17 (11.0)	12 (15.4)	5 (6.6)	0.284
Stillbirth history, n (%)	14 (9.1)	9 (11.5)	5 (6.6)	0.173
Smoking habit, n (%)	27 (17.5)	20 (25.6)	7 (9.2)	0.443
In-vitro fertilization pregnancy, n (%)	2 (1.3)	1 (1.3)	1 (1.3)	0.163
Gravidity, median, IQR^d	1 [0, 3]	2 [1, 3]	1 [0, 2]	−0.149
Hypertension, n (%)	14 (9.1)	6 (7.7)	8 (10.5)	0.098
Diabetes, n (%)	3 (1.9)	1 (1.3)	2 (2.6)	0.097
SLE^e, n (%)	7 (4.5)	2 (2.6)	5 (6.6)	0.193
APS^f, n (%)	3 (1.9)	1 (1.3)	2 (2.6)	0.097
Thrombophilia, n (%)	3 (1.9)	3 (3.8)	0 (0)	0.282
Hypothyroidism, n (%)	24 (15.6)	12 (15.4)	12 (15.8)	0.011

^aACGV: Abdominal circumference growth velocity; ^bSD: standard deviation; ^cBMI: body mass index; ^dIQR: interquartile range; ^eSLE: Systemic lupus erythematosus; ^fAPS: Antiphospholipid syndrome.

Overall, the mean gestational age at birth was 37.76 ± 1.02 weeks, with a mean birthweight of 2,437 g [IQR 2,280, 2,635]. Cesarean delivery was the most frequent delivery mode (73.4%), with no significant differences between groups (Table 2). The presence of a composite APO was relatively common (29.2%) and was higher in the group with abnormal ACGV (41% vs. 17.1%; p = 0.001). Furthermore, intrapartum fetal distress requiring expedited cesarean delivery was also higher in the group with abnormal ACGV (16.7% vs. 5.2%; p = 0.023); elective cesarean delivery rate was 30.1%, with no significant difference between groups (Table 2). There was a significant association between ACGV (i.e. ≤10th centile) and APO (OR 3.37, 95% CI 1.60–7.13; adjusted OR [aOR] 4.30, 95% CI 1.77–10.49) (Tables 3, 4). Likewise, similar effects were observed with a centile of 20 or less for ACGV (Tables 3, 4). Figure 1 shows the predictive performance of ACGV and two other logistic regression models (i.e. EFW and EFW + ACGV) for the presence of a composite APO as continuous variables. Although the AUC for ACGV (0.64; 95% CI 0.54–0.73) was slightly higher compared to that for EFW (0.54; 95% CI 0.44, 0.64) and EFW + ACGV (0.54; 95% CI 0.44, 0.64) for the prediction of APO, it did not reach statistical significance (p = 0.297). There were no perinatal deaths.

Figure 1. The area under the ROC curve of three logistic regression models based on EFW and ACGV expressed as continuous variables for predicting the composite adverse perinatal outcome in small for gestational age fetuses.

Table 2. Perinatal outcomes.

Characteristic	Total N = 154	ACGV^a ≤10th centile N = 78	ACGV >10th centile N = 76	P value
GA^b at delivery, weeks, mean, SD^c	37.76 ± 1.02	37.77 ± 1.03	37.75 ± 1.01	0.920
Birthweight, g, median, IQR^d	2437 [2280, 2635]	2419.5 [2280, 2632]	2469.5 [2282, 2651.5]	0.480
Cesarean delivery, n (%)	113 (73.4)	60 (76.9)	53 (69.7)	0.310
Elective cesarean delivery, n (%)	34 (30.1)	19 (31.7)	15 (28.3)	0.700
IFD^e requiring expedited cesarean delivery, n (%)	17 (11.0)	13 (16.7)	4 (5.2)	0.023
Composite APO^f, n (%)	45 (29.2)	32 (41)	13 (17.1)	0.001

^aACGV: Abdominal circumference growth velocity; ^bGA: gestational age; ^cSD: standard deviation; ^dIQR: interquartile range; ^eIFD: intrapartum fetal distress; ^fAPO: adverse perinatal outcome.

Table 3. Univariable logistic regression models for adverse perinatal outcomes in small for gestational age fetuses.

Odd ratios (OR) (95% CI^a)
ACGV^b centile	ACGV ≤10^th centile	ACGV ≤20^th centile	EFW^c centile
0.98 (0.96, 1.00)	3.37 (1.60, 7.13)	3.25 (1.38, 7.63)	0.94 (0.81, 1.09)

^aCI: confidence interval; ^bACGV: Abdominal circumference growth velocity; ^cEFW: estimated fetal weight.

Table 4. Multivariable logistic regression models for adverse perinatal outcomes in small for gestational age fetuses.

Adjusted odd ratio (aOR^a) (95% CI^b)
ACGV^c centile	ACGV ≤10^th centile	ACGV ≤20^th centile	EFW^d centile
0.98 (0.95, 1.00)	4.30 (1.77, 10.49)	3.86 (1.42, 10.55)	0.93 (0.79, 1.10)

^aaOR: adjusted effect for maternal age, pregestational BMI, preeclampsia history, SGA history, fetal demise history, tobacco exposure, conception method, hypertension, diabetes, lupus, antiphospholipid syndrome, thrombophilia, hypothyroidism; ^b CI: confidence interval; ^cACGV: Abdominal circumference growth velocity; ^d EFW: estimated fetal weight.

Discussion

The results of the present study show that despite being a population in which very low morbidity would be expected, SGA fetuses without clinical evidence of placental insufficiency (i.e. those considered constitutionally small) present a high risk of APO. Almost one-third of our cohort (29.2%) had at least one APO. This finding highlights the importance of identifying those fetuses at risk of APO within the population of constitutionally small fetuses.

Identifying SGA fetuses at increased risk of perinatal mortality and adverse neurodevelopmental outcomes (i.e. growth-restricted fetuses) is only sometimes straightforward. Longitudinal fetal growth assessment has recently been considered a tool that could aid in identifying those fetuses failing to reach their growth potential [6,9]. We demonstrated an association between an abnormal ACGV and APO in a homogeneous cohort of SGA fetuses without clinical evidence of placental insufficiency. Likewise, when intrapartum fetal distress was analyzed independently as an indication for cesarean section, a significant association was found with an abnormal ACGV; this is a relevant finding, as intrapartum fetal distress is a variable that reflects the consequences of impaired fetal growth [22].

The rationale for using AC growth as a proxy of longitudinal fetal growth is twofold. First, AC has been described to have a relatively steady growth pattern across gestational age as opposed to other skeletal parameters (e.g. head circumference, biparietal diameter, femur length) [9]. Second, AC growth is strongly influenced by the fetus underlying nutritional status [23]. There are several methods to calculate growth velocity [9,20]. The formula we used to calculate ACGV uses the Z score difference between two ultrasound examinations. Both the Z score and centile position are expressions of fetal size; however, the significance of changing or crossing centiles depends on the position along the normal distribution, while the Z score calculation formula can overcome this limitation. We chose to analyze pairs of observations taken during the second and third trimesters, as it has been shown that larger time intervals between measurements are associated with less variability in growth and measurement error [10,23]. Moreover, both time frames are common gestational windows for routine scans (i.e. anomaly and growth scans). Analyzing ACGV as a dichotomic variable (i.e. ≤10th centile and >10th centile) implies a potential overlap in the study population; the abnormal ACGV group could include some normal yet constitutionally small fetuses and vice versa; the ratio of constitutionally small fetuses to fetuses whose growth potential has been compromised will depend on the prevalence of such illnesses in a given population. However, using an ACGV cutoff centile turns a complex measure into an intuitively more understandable tool for clinical practice.

Our results align with those reported by other investigators in different populations. Hendrix et al. in a study of fetuses whose birthweight was between the 10–80th centile, observed that newborns with APO had significantly lower ACGV (p = 0.03) [24]. Moreover, Kennedy et al. studied the relationship between ACGV and other indirect markers of placental insufficiency in a population of adequate weight for gestational age fetuses, observing that the odds of having an umbilical artery pH <7.15, a small placenta (<10th centile) and an abnormally low body fat percentage increased for each centile drop in the ACGV [25]. Likewise, in a prospective cohort in which a universal third-trimester scan was performed, an EFW <10th centile was significantly associated with APO (risk ratio [RR] 1.6; 95% CI 1.2–2.1); however, when ACGV was included in the analysis, the risk of APO increased only in those fetuses that in addition to having an EFW <10th centile also had an altered ACGV (RR, 3.9; 95% CI 1.9–8.1). [20] Cavallaro et al. who have already investigated the relationship between fetal ACGV and APO in SGA fetuses, concluded that the main benefit of using ACGV is the decrease in false positive results [26]. However, it should be noted that this cohort included fetuses with EFW below the third centile and clear evidence of placental insufficiency (i.e. abnormal Doppler evaluation), which should be considered growth-restricted fetuses.

To the best of our knowledge, this is one of the first studies to analyze the ACGV in a homogeneous population of SGA fetuses without clear evidence of placental insufficiency or adaptation to hypoxia, in which the scans were performed by maternal-fetal medicine specialists, according to standardized guidelines [16]. Our findings could have important implications for clinical practice, as using ACGV in SGA fetuses without evidence of placental insufficiency could potentially lead to identifying those with disturbed growth. However, we acknowledge the limitations of our study. First, given the retrospective nature of the study, the outcomes were subject to intervention bias, as reflected by the cesarean delivery rate and the mean gestational age at birth. The latter is of utmost importance as some of the APOs of interest can be related to both the condition (i.e. placental insufficiency resulting in hypoxia) and its treatment (i.e. relatively premature delivery). Second, the sample size hampered the individual analysis of the different APOs due to their low prevalence; nevertheless, studying them as a composite outcome facilitates the analysis and provides valuable information to guide clinical care and to plan future studies. Third, insults at distinct time points during pregnancy are expected to have different effects on fetal growth and development [23]. This assumption is a source of variability in ACGV and, therefore, could partially account for the AUC values not reaching statistical significance. Finally, there are practical challenges to adopting ACGV assessment in routine clinical practice, as measurements could be complex and time-consuming; however, using easy-to-use and free online calculators such as the one used in this study could facilitate the process. Large-scale and multicenter studies will be needed to confirm our results and use ACGV as a clinical tool for evaluating and surveilling SGA fetuses.

Conclusion

Our results suggest that an abdominal circumference growth velocity below the 10th centile is a risk factor for adverse perinatal outcomes among small for gestational age fetuses.

Acknowledgments

The authors wish to thank the staff of the Maternal-Fetal Medicine Department at the National Perinatology Institute, where this study was conducted, for their support and guidance.

Disclosure statement

The authors declare that they have no known competing financial interests or personal relationships that could have influenced the work reported in this manuscript.

Data availability statement

The data that support the findings of this study are available from the corresponding author, [MJ-RS], upon reasonable request.

Additional information

Funding

No funding was obtained for the present study.