Globally, including in South Korea, the incidence of gestational diabetes and diabetes (DM) during pregnancy is increasing annually owing to factors such as aging of the childbearing population and an increase in maternal obesity rates. The prevalence of gestational diabetes in South Korea has increased from 15.8% in 2017 to 18.2% in 2021, with a rate of 22.5% among pregnant women aged >40 years [1].

Mothers with gestational diabetes mellitus (GDM) or pregestational diabetes mellitus (PGDM) may experience various obstetric complications including preeclampsia, preterm labor, and polyhydramnios [2,3]. The risk of prenatal and neonatal complications is also elevated in these mothers, such as fetal hyperglycemia, hyperinsulinemia, and fetal overgrowth (e.g., macrosomia and large for gestational age [LGA]), which may occur as a result [4,5] and is significantly associated with increased rates of perineal injury, shoulder dystocia, birth injury, and cesarean section [5]. In the long term, persistent insulin increments in the fetus can cause neonatal hypoglycemia at birth and are associated with neurological sequelae, such as developmental delays, seizures, cognitive difficulties, and the risk of glucose metabolism disorders. In addition, the obesity rate was high in children born to mothers with GDM and PGDM mothers [6-11].

During maternal hyperglycemia, glucose passes through the placenta to the fetus, causing fetal hyperglycemia. Because insulin does not pass through the placenta, high fetal glucose levels cause the fetal pancreas to produce much insulin, resulting in hyperinsulinemia. Fetal hyperinsulinemia stimulates the anabolic pathway, which causes overgrowth of connective tissue, fat tissue, and muscles, ultimately resulting in fetal overgrowth [9].

Fetal ultrasound measurements in mothers with GDM are crucial for predicting and preventing maternal and perinatal complications. Several studies have been conducted in this area. The abdominal circumference (AC) of the fetus, a major factor associated with fetal weight, has been identified in many studies as a predictor of macrosomia and LGA status [12-18]. Macrosomic infants of mothers with DM tend to have more overall body fat and a smaller head circumference (HC)/AC ratio, often due to excessive insulin secretion, which causes organomegaly and disproportionate fat accumulation, causing the torso to outgrow the head [19-21].

To date, few studies have investigated the direct relationship between fetal biometry and neonatal outcomes. There is a paucity of studies on pregnant women in South Korea. Therefore, in this study, based on previous research, we aimed to investigate the relationship between the abdominal and HCs of fetuses and the incidence of neonatal complications in mothers with DM compared with normal pregnancies.

This retrospective study included expectant mothers with DM (GDM or PGDM) and those with normal pregnancy. Data were retrieved from hospital registers focusing on pregnant women who were diagnosed with GDM or PGDM according to national guidelines [22] and who gave birth at a single tertiary center between January 2013 and December 2022.

The inclusion criteria were as follows: age ≥18 years, singleton pregnancy, and delivery of a full-term baby (≥37+0 weeks of gestation).

We excluded patients with no ultrasound results approximately 2 weeks before delivery, women whose newborns could not undergo workup, and women diagnosed with gestational hypertensive disorders (such as preeclampsia and eclampsia) in an otherwise normal pregnancy.

Data on maternal age, weight, body mass index (BMI), and parity were collected, and treatment with dietary advice or insulin was assessed in mothers with DM. Data from all mothers with DM were analyzed and categorized into GDM and PGDM groups. The modes of delivery included spontaneous delivery, labor induction, and cesarean section.

For fetal biometric measurements, the ultrasound-measured estimated fetal weight (EFW), fetal HC, AC, difference between the head and abdominal circumference (AC-HC), and HC/AC ratio were determined. EFW was calculated using the Hadlock IV formula, which combines the measurements of biparietal diameter (BPD), HC, AC, and femur length (FL). BPD, HC, AC, and FL measurements followed the guidelines presented by the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) [23]. Cut-off values for AC-HC and HC/AC ratio were set to ≥2.5 and 0.95, respectively [2,21].

Neonatal outcomes were analyzed by assessing neonatal intensive care unit (NICU) admission, intubation, and hypoglycemia, using data retrieved from the Haeundae Paik Hospital registers [24]. Hypoglycemia was defined as a blood glucose level <50 mg/dL, based on the definitions of the American Academy of Pediatrics and Pediatrics Endocrine Society [4].

All statistical analyses were performed using SPSS version 26.0 for Windows (IBM Corp., Armonk, NY, USA) and R 4.1.2 (R Core Team, R Foundation for Statistical Computing, Vienna, Austria).

Variables were presented as frequencies and percentages for categorical data and mean values±standard deviations for quantitative data. Group differences were tested using the chi-square test or Fisher’s exact test for categorical data and the independent t-test or Mann-Whitney U-test for quantitative data, as appropriate. The Shapiro-Wilk test was used to check the normality of the quantitative data distribution.

Univariate and multivariate analyses using binary logistic regression were conducted to identify the factors independently associated with the outcome variables.

This study received Institutional Review Board (IRB) approval, and the requirement for informed consent was waived because of the retrospective nature of the study (IRB No. HPIRB2023-05-040-001).

Of the 473 pregnant women, 175 (37.0%) had diabetic and 298 (63.0%) had normal pregnancies. Of all the other DM cases, 138 (78.9%) had GDM and 37 (21.1%) had PGDM. The characteristics of the candidates are listed in Table 1.

The mean gestational ages at delivery were 38 weeks and 2 days, respectively. The mean age was 34.7±4.4 years, and individuals with DM overall had a high average age compared to those without (35.5±4.5 vs. 34.2±4.3). The BMI at delivery was 29.5 kg/m² in overall mothers with DM (28.9±5.7 in GDM mothers, 31.9±6.0 in PGDM mothers) and 26.5 kg/m² in healthy mothers, which was significantly higher than that in mothers with DM. The BMI before pregnancy was also higher in mothers with DM than in healthy mothers. The rate of vaginal delivery in mothers with DM was 29.1%, whereas 70.9% of births occurred via cesarean section. In healthy mothers, the rates were 35.9% and 64.1%, respectively, with no significant differences between the two groups.

The median EFW at the time of delivery was 3,048.4, 2,997.7, 3,237.5, and 3,017.8 g for overall diabetic, GDM, PGDM, and healthy mothers, respectively, whereas the birth weights were 3,176.7, 3,115.2, 3,405.9 and 3,107.4 g, respectively. Among the percentile birth weights, the proportion of fetuses confirmed to be LGA was 12.6% in mothers with DM, 10.9% in mothers with GDM, 18.9% in mothers with PGDM, and 2.2% in healthy mothers, with the difference between the values being statistically significant.

Multivariate analyses were performed on the association between NICU admission, intubation, hypoglycemia, and fetal ultrasound measurements at delivery (HC/AC <0.95 or AC-HC ≥2.5 cm) as indicators of neonatal prognosis.

In mothers diagnosed with overall DM, HC/AC was <0.95, and significantly associated with neonatal hypoglycemia (P<0.05), and the odds ratio (OR) was 2.785 (95% confidence interval [CI], 1.335-5.808). When mothers with DM were divided into two groups for analysis, the GDM group demonstrated no statistically significant association (P=0.29); however, the PGDM group showed a significant correlation (P<0.05; OR, 24.29; 95% CI, 2.83-208.12). Neonatal hypoglycemia in healthy mothers was not significantly associated with HC/AC (P=0.91). Additionally, there was no statistically significant association among NICU admission, intubation, and HC/AC in any group (Fig. 1).

When AC-HC was ≥2.5 cm, the OR for neonatal hypoglycemia in overall mothers with DM was 2.451 (95% CI, 1.041-5.770), indicating a statistically significant association (P<0.05). However, when categorized into GDM and PGDM, a statistically significant association was observed only in the PGDM group (P<0.05; OR, 11.33; 95% CI, 1.80-71.51) but not in GDM group (P=0.83). In healthy mothers, there was no significant association between neonatal hypoglycemia and AC-HC (P=0.75). Other prognostic factors were not significantly associated with AC-HC in DM or healthy mothers. There were also no statistically significant associations between NICU admission, intubation, and AC-HC in the two groups (Fig. 2).

In addition to the increase in obesity rates, the proportion of people diagnosed with DM in Korea is gradually increasing due to routine health checkups. The age at the first diagnosis of DM is decreasing; consequently, the number of pregnancies with DM and the rate of gestational diabetes diagnosis are gradually increasing.

The complications that can arise from GDM and PGDM are diverse. Fetuses can develop macrosomia, shoulder dystocia, low APGAR scores at birth, and neonatal hypoglycemia [2,25].

Several studies have explored the relationship between various complications occurring in maternal DM and fetal ultrasound measurements. We aimed to determine whether neonatal complications could be predicted through fetal ultrasound measurements in mothers with DM based on the characteristic growth pattern of the fetus and cut-off values from previous studies.

In mothers with DM, excessive insulin secretion can lead to improper fat accumulation and disproportionate fetal growth, that is, the body grows more than the head. Duewel et al. [21] quantified this imbalance using AC-HC and HC/AC parameters. They explored whether these metrics could predict shoulder dystocia, a common complication in mothers with DM, in a large cohort study. However, their study was conducted in Germany; thus, it does not represent diverse ethnic groups, including Koreans, limiting its direct application to our population. Nonetheless, it presents fetal biometry as a clear and specific risk factor. They proposed an antenatal risk scoring system for predicting shoulder dystocia, with DM, AC-HC, and EFW as independent predictors. Among these, an AC-HC of ≥2.5 cm was reported as the cut-off value signifying asymmetrical growth. Based on this, we considered the HC/AC to reflect disproportionate growth in the fetuses of mothers with DM and incorporated the 0.95 cut-off value suggested in our previous study.

We found that the rate of HC/AC <0.95 or AC-HC ≥2.5 was higher in overall mothers with DM than in healthy mothers, which was statistically significant (P<0.05). Additionally, we identified a significant association between HC/AC <0.95 or AC-HC ≥2.5 and neonatal hypoglycemia. This association could be explained by increased fetal insulin secretion due to maternal hyperglycemia, exacerbating the imbalance in the growth of HC and AC and worsening neonatal hypoglycemia. When mothers with DM were divided into GDM and PGDM groups, a statistically significant association with neonatal hypoglycemia was observed only in the PGDM group when HC/AC was <0.95 or AC-HC was ≥2.5 cm. However, the 95% CI was wide. Therefore, the reliability of this statistical result was low, possibly because the population size of the group was small. Additionally, the use of insulin and blood glucose controls in each group may have influenced the results. In the data analysis, most mothers with GDM controlled their blood glucose levels through dietary modifications, suggesting that their blood glucose levels did not increase or fluctuate to an extent that would require insulin use. Conversely, most mothers with PGDM used insulin, but given that their insulin sensitivity and resistance were greater than those of mothers with GDM [26], maintaining adequate blood glucose levels may be more challenging, even with insulin use. However, the degree of blood glucose control related to insulin use could not be confirmed because of a lack of data.

Various studies have investigated the associations between maternal DM and neonatal respiratory morbidity, low APGAR scores, and cord acidosis. Previous research has indicated that cord blood acidosis and low APGAR scores occur more frequently in diabetic pregnancies than in nondiabetic pregnancies [27,28]. According to Kawakita et al. [29], respiratory distress syndrome, transient tachypnea of the newborn, and mechanical ventilation are more common in full-term mothers with DM, suggesting a strong association between DM and respiratory morbidity. Although the mechanism is not certain, there is a theory that hyperglycemia-induced hyperinsulinemia slows lung maturation [29,30]. Another study found an association between maternal hyperglycemia within 6 hours of delivery and perinatal asphyxia. However, fetal macrosomia generally does not act as a risk factor for neonatal asphyxia due to cesarean delivery [27].

In a study on fetal ultrasound measurements and neonatal prognosis, there was a study showing that neonatal complications were high in AC ≥75th percentile, and an association between adverse perinatal outcomes and pregnancies with GDM or PGDM, particularly when late pregnancy EFW is above the 90th percentile and the Doppler shows a low cerebro-placental ratio [31,32].

Our study had several limitations. First, we did not identify an association between insulin use and non-use among mothers with DM and neonatal outcomes. As mentioned earlier, the lack of data on blood glucose control makes it impossible to examine the relationship between insulin use and neonatal outcomes. Although our data support the hypothesis that the relationship between asymmetrical fetal growth and neonatal hypoglycemia may be exacerbated by maternal hyperglycemia, there may be limitations because the relationship between fetal biometry and neonatal hypoglycemia was not analyzed depending on whether plasma glucose was controlled. Maternal hyperglycemia is reportedly associated with fetal macrosomia, preterm birth, and LGA. It has also been confirmed that appropriate blood glucose control reduces neonatal hypoglycemia and consequential neonatal complications [33-35]. Additionally, appropriate control is likely to reduce disproportionate fetal growth, which is also related to fetal ultrasound measurements.

Second, patients with hypertensive diseases such as preeclampsia were not excluded from the mother with DM group. This means that other factors (in addition to DM) that can affect fetal growth, such as growth restriction, have not been excluded.

Third, there may be errors due to the timing of fetal ultrasound measurements in relation to delivery. Fetal ultrasound measurements were conducted from 2 weeks before delivery to the day of delivery, and variations in measurement timing may have influenced the results. Moreover, there is no established standard for neonatal hypoglycemia in Korea, leading to various references across institutions and potential reproducibility issues. In addition, the lack of clear NICU admission guidelines at our institution may have affected our results.

Finally, the sample size is small. Particularly, when mothers with DM were divided into GDM and PGDM groups, the P-value indicated statistical significance, but a wide CI was observed.

Despite these limitations, this study provides valuable insights into the association between fetal biometry measurements and neonatal outcomes in the context of DM in pregnancy.

To the best of our knowledge, this is the first study conducted in Korea to confirm the relationship between fetal ultrasound biometric measurements and subdivided neonatal outcomes, especially hypoglycemia, in pregnancies with DM. Furthermore, ultrasound measurements were consistently obtained by an expert at a single center according to the ISUOG guidelines, which means that measurement errors are expected to be minimized.

In this study, we investigated the relationship between fetal biometric measurements in pregnant women with DM and the occurrence of neonatal complications, specifically neonatal hypoglycemia. Our findings provide significant evidence of an association between fetal biometry measurements in diabetic pregnancies and neonatal complications, especially in Korean mothers. Future large-scale, multicenter studies should be conducted to develop non-invasive methods of predicting neonatal outcomes using characteristic fetal biometry measurements in mothers with DM and confirm the cutoff values of HC/AC and AC-HC applicable to Koreans for use in prenatal counseling and maternal care.