2023 Volume 70 Issue 5 Pages 511-517
Hypoglycemia is one of the most significant problems in neonates born to mothers with gestational diabetes (GDM). This study aimed to identify novel predictors of hypoglycemia in neonates born to mothers with GDM. A total of 443 term singleton infants from mothers diagnosed with GDM and cared for at Keio University Hospital between January 2013 and December 2019 were included in this study. Neonatal hypoglycemia was defined as hypoglycemia of less than 47 mg/dL at 1 or 2 or 4 h after birth, according to previous studies. Among 443 full-term singleton neonates born to mothers with GDM, 200 developed hypoglycemia (45%). Gestational weight gain (GWG), HbA1c at 1st trimester, HbA1c at GDM diagnosis, and the incidence of insulin therapy in the neonatal hypoglycemia group were significantly higher than those in the non-neonatal hypoglycemia group (p = 0.016, p = 0.032, p = 0.011, and p = 0.017, respectively). Regarding the multiple regression analysis adjusted for nulliparity, GWG, and gestational weeks at delivery, the odds ratio for maternal HbA1c ≥5.2% at 1st trimester was 1.63 (p = 0.034), and maternal insulin therapy during pregnancy was 1.72 (p = 0.015). In conclusion, HbA1c in the 1st trimester and insulin therapy during pregnancy were good predictors of hypoglycemia in neonates born to GDM mothers, especially when their HbA1c was 5.2% or more. Further research will be necessary to improve the perinatal management of hypoglycemia.
GESTATIONAL DIABETES MELLITUS (GDM) is associated with cesarean section (CS) delivery, shoulder dystocia, macrosomia (birth weight >4,000 g), neonatal hypoglycemia and jaundice, obesity, and metabolic syndrome in the next generation [1-3]. Because maternal glucose passes through the placenta and fetuses born to GDM mothers develop hyperinsulinemia, neonates born to mothers with GDM often present with hypoglycemia, although infants with hypoglycemia can be born to women with well-controlled GDM [4].
Hypoglycemia is an important neonatal complication associated with irreversible brain damage, and previous studies have reported an association between neonatal hypoglycemia, brain damage, and neurodevelopmental disorders [5, 6]. The severity of hypoglycemia is important in determining the neurodevelopmental prognosis of neonates with hyperinsulinemic hypoglycemia [7, 8]. This indicates that maternal hyperglycemia in early pregnancy (before GDM diagnosis) may contribute to neonatal hypoglycemia and adversely affect fetal tissue development. In the long-term, although neonatal hypoglycemia was not associated with increased risk of combined neurosensory impairment, it was associated with a dose-dependent increased risk of poor executive function and visual motor function in children at 4.5 years of age [9].
Furthermore, hypoglycemia in newborns predicts future development of obesity and metabolic syndrome [10]. For example, it is known that a high maternal dietary glycemic index and glycemic load in early pregnancy can lead to a larger fetal abdominal circumference in late pregnancy and higher fat mass in the offspring even at 4–6 years of age [11, 12]. Fluctuations in glucose and insulin concentrations in early pregnancy have been associated with glucose and insulin concentrations during childhood [13]. Therefore, it is important to predict and prevent neonatal hypoglycemia based on maternal or perinatal information during pregnancy [e.g., maternal body mass index (BMI) and oral glucose tolerance test (OGTT)]. However, there is a paucity of data on maternal or perinatal information for predicting hypoglycemia in neonates born to mothers with GDM.
This study aimed to identify novel predictors of hypoglycemia in neonates born to mothers with GDM.
With the approval of the Keio University Ethics Committee (Nos. 20150103 and 20150168), a total of 443 term singleton infants from mothers who were diagnosed with GDM and cared for at Keio University Hospital between January 2013 and December 2019 were included. We excluded women with multi-fetal pregnancies, fetal congenital anomalies, overt diabetes during pregnancy, and pre-existing diabetes (either type 1 diabetes or type 2 diabetes). GDM was diagnosed according to the Japan Society of Obstetrics and Gynecology (JSOG) criteria, modified by the International Association of Diabetes in Pregnancy Study Group (IADPSG) criteria (i.e., the World Health Organization 2013 criteria) using the 75 g oral glucose tolerance test (OGTT) [14]. In our hospital, the 1st trimester screening for GDM was performed as described in our previous report (e.g., pre-pregnancy BMI ≥25 kg/m2, random plasma glucose level ≥95 mg/dL, HbA1c at the 1st trimester ≥5.9%, prior GDM, GA ≥15.8%, a personal history of GDM or macrosomia delivery, and family history of type 2 DM) [15]. If the 1st trimester screening for GDM was negative or positive but OGTT was negative, pregnant women had to undergo the 50 g glucose challenge test (GCT) at about 24 gestational weeks. If the patients were GCT positive, they underwent an OGTT at 24–28 gestational weeks. The threshold of OGTT in both GDM before and after 24 gestational weeks were both as follows: fasting plasma glucose level (PG) ≥92 mg/dL, 1h-PG ≥180 mg/dL, and 2h-PG ≥153 mg/dL. As described in our previous report, all women with GDM received dietary counseling before beginning a dietary regimen of three meals and three snacks daily. Capillary glucose profiles were obtained via self-monitored blood glucose (SMBG) measurements performed seven times per day: once upon waking up, one hour prior to all planned meals, two hours after each meal, and once at bedtime. Dietary management was based on the recommended daily caloric intake for each individual. Standard body weight (SBW), measured in kilograms, was calculated as maternal height (m) × maternal height (m) × 22. This value was then substituted into one of the three equations based on body condition score and stage of pregnancy: early pregnancy, 30 kcal × SBW + 150 kcal; late pregnancy, 30 kcal × SBW + 350 kcal; and if obese, 30 kcal × SBW throughout pregnancy. Insulin was administered when dietary treatment alone did not achieve the glycemic goal (i.e., FPG level <100 mg/dL or 2 h PG <120 mg/dL). GDM patients with one positive point on the diagnostic OGTT were managed as outpatients, while GDM patients with two or three positive points were first hospitalized for one week for education on GDM management and then managed as outpatients [15]. Maternal and neonatal data were collected from medical records. Neonatal hypoglycemia was defined as blood glucose of less than 47 mg/dL at 1 or 2 or 4 h after birth, according to previous studies [4]. Hyperglycemia during labor is defined as a PG level of ≥126 mg/dL measured at the hospital within 24 hours of delivery [16]. Since this was a retrospective study, we received a waiver for obtaining informed consent from patients.
Data are presented as the median (range) or number of cases (percentage). Continuous data were compared between the groups using the Mann–Whitney U test. Categorical variables were analyzed using the chi-squared or Fisher’s exact test. Prepregnancy BMI and gestational weight gain (GWG) trends were analyzed using Cochran–Armitage trend analysis. In all tests, p < 0.05 was considered statistically significant. The predictive accuracy of HbA1c in the 1st trimester was assessed using the area under the receiver operating characteristic curve (ROC-AUC) and the calculated cutoff value to predict neonatal hypoglycemia. Multiple logistic regression analysis was performed to evaluate the relative contributions of the various maternal metabolic management factors to neonatal hypoglycemia, adjusted for nulliparity, GWG, and gestational weeks at delivery. The following independent variables were included in the model based on the clinical relevance and statistical significance of their association with neonatal hypoglycemia: HbA1c level in the 1st trimester and insulin therapy during pregnancy. Neonatal BMI at one month was calculated as weight (kg)/height2 (m2), and ponderal index (PI) was calculated as weight × 100/height3 (cm3) [17]. Odds ratios (OR) and 95% confidence intervals (CI) were evaluated to determine associations between maternal metabolic features and neonatal hypoglycemia. Statistical analyses were performed using JMP software (ver. 15, SAS Inst. Inc., Cary, NC, USA).
In this study, among 443 full-term singleton neonates born to mothers with GDM, 200 (45%) neonates developed hypoglycemia. Comparisons of maternal and neonatal characteristics between the neonatal hypoglycemia and non-hypoglycemia groups are shown in Table 1. The incidence of nulliparity and gestational weeks at delivery in the neonatal hypoglycemia group was significantly lower than that in the non-neonatal hypoglycemia group (p = 0.0081 and p = 0.0002, respectively). GWG, HbA1c at 1st trimester, HbA1c at GDM diagnosis, and the incidence of insulin therapy in the neonatal hypoglycemia group were significantly higher than those in the non-neonatal hypoglycemia group (p = 0.016, p = 0.032, p = 0.011, and p = 0.017, respectively). However, the two groups had no differences in glycated-albumin (GA), glucose levels during OGTT, number of positive points, PG during labor, or birth weight of neonates. There was no difference in neonatal weight, BMI, or PI between the two groups at one month.
| Neonatal hypoglycemia (n = 200) | Non-neonatal hypoglycemia (n = 243) | p-value | ||
|---|---|---|---|---|
| Maternal age at delivery | (years) | 37 (25–57) | 37 (23–59) | 0.53 |
| Nulliparity | 109 (55%) | 163 (67%) | 0.0081 | |
| Pre-pregnancy BMI | (kg/m2) | 21 (16.4–41.1) | 21.4 (17.0–36.3) | 0.39 |
| Obesity (≥25 kg/m2) | 26 (13%) | 42 (17%) | 0.23 | |
| Obesity (≥30 kg/m2) | 10 (5%) | 12 (5%) | 1 | |
| Gestational weight gain | (kg/40weeks) | 9.3 (–2.3–20.5) | 8.5 (–8.6–19.9) | 0.016 |
| Gestational weight gain category | 0.0021 | |||
| Inadequate | 117 (59%) | 173 (71%) | ||
| Appropriate | 57 (29%) | 40 (16%) | ||
| Excessive | 16 (8%) | 21 (9%) | ||
| GDM diagnosed before 24 gestational weeks | 110 (55%) | 139 (57%) | 0.7 | |
| HbA1c | (%) | |||
| At the 1st trimester | 5.4 (4.8–7.0) | 5.3 (3.6–6.2) | 0.032 | |
| At diagnosed GDM | 5.2 (4.5–6.2) | 5.2 (3.4–6.1) | 0.011 | |
| At the 3rd trimester | 5.4 (4.6–6.4) | 5.5 (3.6–6.3) | 0.96 | |
| Glycated-albumin | ||||
| At the 1st trimester | 13.8 (10.3–17.7) | 13.8 (10.2–16.8) | 0.34 | |
| At diagnosed GDM | 13.3 (9.6–16.7) | 13.1 (10.2–16.1) | 0.19 | |
| At the 3rd trimester | 12.7 (9.6–15.9) | 12.6 (10.1–15.7) | 0.32 | |
| Results of 75 g OGTT | ||||
| Fasting glucose level | (mg/dL) | 91 (54–114) | 92 (58–118) | 0.22 |
| 1-hour glucose level | (mg/dL) | 175 (89–272) | 171 (71–263) | 0.32 |
| 2-hour glucose level | (mg/dL) | 154 (87–292) | 149 (81–235) | 0.075 |
| Fasting glucose level positive | 97 (49%) | 132 (54%) | 0.25 | |
| 1-hour glucose level positive | 91 (46%) | 94 (39%) | 0.18 | |
| 2-hour glucose level positive | 107 (54%) | 111 (46%) | 0.11 | |
| Number of positive | 0.15 | |||
| 1 point positive | 122 (61%) | 171 (70%) | ||
| 2 point positive | 61 (31%) | 50 (21%) | ||
| 3 point positive | 17 (9%) | 22 (9%) | ||
| Insulin use during pregnancy | 85 (43%) | 76 (31%) | 0.017 | |
| Unit of insulin at delivery | (unit) | 12 (2–54) | 16.5 (4–65) | 0.12 |
| Hyperglycemia during delivery | 73 (38%) | 84 (37%) | 0.84 | |
| Unknown | 10 | 10 | ||
| Gestational weeks at delivery | (week) | 38 (37–41) | 39 (37–41) | 0.0002 |
| Neonatal sex (female) | 95 (48%) | 123 (51%) | 0.57 | |
| Birthweight | (g) | 2,925 (2,206–4,526) | 3,014 (2,054–4,042) | 0.091 |
| Low birthweight (<2,500 g) | 18 (9%) | 22 (9%) | 1 | |
| Small for gestational age (<10% tile) | 14 (7%) | 18 (7%) | 1 | |
| Large for gestational age (≥90% tile) | 30 (15%) | 34 (14%) | 0.79 | |
| Macrosomia | 2 (1%) | 1 (0%) | 0.59 | |
| Neonatal weight at one month | (g) | 4,144 (2,500–5,640) | 4,130 (2,918–5,806) | 0.53 |
| BMI | (kg/m2) | 14.6 (11.6–17.6) | 14.7 (11.4–19.0) | 0.98 |
| PI | (g × 100/cm3) | 2.7 (2.2–3.5) | 2.7 (2.1–3.8) | 0.38 |
BMI: body mass index, OGTT: oral glucose tolerance test, GDM: gestational diabetes mellitus, PI: ponderal index. Data was median (range) or n (%).
According to the ROC curve using HbA1c in the 1st trimester, the AUC was 0.56 (p = 0.0093, sensitivity: 80%, and specificity: 33%) (Fig. 1). The cutoff value of HbA1c in the 1st trimester for predicting neonatal hypoglycemia was 5.2%. Regarding the multiple regression analysis adjusted for nulliparity, GWG, and gestational weeks at delivery, the odds ratio for maternal HbA1c ≥5.2% at 1st trimester was 1.63 (p = 0.034), and maternal insulin therapy during pregnancy was 1.72 (p = 0.015) (Table 2).
Receiver-operating characteristics (ROC) curve analysis for the prediction of neonatal hypoglycemia for HbA1c in the 1st trimester. A cut-off value of 5.2% had a sensitivity of 80% and specificity of 33%; area under curve, 0.56.
| Univariate | Multiple | ||||||
|---|---|---|---|---|---|---|---|
| Odds ratio | 95% CI | p-value | Odds ratio | 95% CI | p-value | ||
| HbA1c at the 1st trimester ≥5.2% | No | 1 (reference) | 1 (reference) | ||||
| Yes | 1.73 | 1.08–2.79 | 0.023 | 1.68 | 1.04–2.70 | 0.034 | |
| Insulin use during pregnancy | No | 1 (reference) | 1 (reference) | ||||
| Yes | 1.71 | 1.13–2.60 | 0.011 | 1.72 | 1.11–2.66 | 0.015 | |
CI, confidence interval.
This study aimed to identify novel predictors of hypoglycemia in neonates born to mothers with GDM. Approximately half of the term singleton neonates born to mothers with GDM developed hypoglycemia. HbA1c in the 1st trimester, but not in the 3rd trimester, and insulin use during pregnancy might be among the most important predictors of neonatal hypoglycemia, and the cutoff value of HbA1c in the 1st trimester to predict neonatal hypoglycemia was 5.2%.
In the present study, approximately half of the term singleton neonates born to mothers with GDM developed hypoglycemia. This incidence was higher than that reported in previous studies [18-20]. However, neonatal hypoglycemia had several definitions. The present study referenced its definition of neonatal hypoglycemia from the report by Voormolen et al., and the incidence was the same as that reported in the present study (51%) [4]. Furthermore, since high-risk infants (e.g., large for gestational age) occur more frequently than appropriate for gestational age (71%), the difference in participants may influence the result. HbA1c is an important marker for the diagnosis and management of maternal glucose tolerance. It was previously reported that HbA1c level in the 3rd trimester was a good predictor of neonatal hypoglycemia. For example, Mikkelsen et al. reported that high maternal HbA1c levels were associated with neonatal hypoglycemia and an overweight BMI [21]. Particularly, HbA1c levels ≥5.5% are associated with neonatal hypoglycemia [22]. In another report, HbA1c level >5.2% in the 3rd trimester was associated with neonatal hypoglycemia, especially in preterm birth [23]. While HbA1c in the 3rd trimester was not associated with neonatal hypoglycemia in this study, HbA1c in the 1st trimester and GDM diagnosis were associated with neonatal hypoglycemia. However, there have been no previous reports that high HbA1c levels in the 1st trimester are associated with neonatal hypoglycemia. High-risk women with GDM (e.g., pre-pregnancy BMI ≥25 kg/m2, random plasma glucose level ≥95 mg/dL, HbA1c at the 1st trimester ≥5.9%, prior GDM, GA ≥15.8%, a personal history of GDM or macrosomia delivery, and family history of type 2 DM) underwent OGTT in the 1st trimester and were diagnosed before 24 gestational weeks in this study [15]. The Japanese diagnostic criteria differ from some foreign criteria, and high-risk women are diagnosed and managed for GDM from early pregnancy. Therefore, since many GDM mothers were well-controlled by the 3rd trimester, there might not be a difference in HbA1c in the 3rd trimester between the two groups. Contrastingly, early onset GDM was associated with a higher risk of neonatal hypoglycemia than late-onset GDM in a meta-analysis [24]. However, the incidence of GDM diagnosed before 24 weeks of gestation in the neonatal hypoglycemia group was not different from the non-neonatal hypoglycemia group in this study. We previously reported that umbilical blood DNA methylation in GDM diagnosed before 24 gestational weeks was not different from that in GDM diagnosed after 24 gestational weeks and a normal OGTT [25]. Therefore, glucose control to prevent changes in umbilical blood DNA methylation could prevent neonatal hypoglycemia. Furthermore, although higher GA during late pregnancy was associated with neonatal hypoglycemia in a small sample size analysis of Japanese mothers with diabetes and GDM [26, 27], GA was not associated with neonatal hypoglycemia in this study.
Although it has been reported that there is no difference in neonatal hypoglycemia between insulin-treated and diet-controlled GDM mothers among Caucasian [4, 28], insulin use during pregnancy was associated with neonatal hypoglycemia in this study. Moreover, in Taiwanese newborns born after 35 gestational weeks, insulin treatment during pregnancy increased the risk of neonatal hypoglycemia [18]. Another study reported that insulin therapy did not increase the risk of neonatal hypoglycemia compared to oral hypoglycemic agents [29]. Unfortunately, since no antidiabetic pharmacological therapies have been approved for the treatment of GDM in Japan, we could not analyze the association between antidiabetic pharmacological therapies and neonatal hypoglycemia. Furthermore, since the diagnostic criteria of GDM are not the same as those in other countries, the differences in diagnostic criteria of GDM may have influenced the association between insulin treatment and neonatal hypoglycemia. Although hyperglycemia during labor is one of the most critical risk factors of neonatal hypoglycemia [30], it was not associated with neonatal hypoglycemia in the present study. The patients were in several situations; for example, some delivered vaginally, whereas others delivered via CS. Furthermore, because some patients delivered within 24 hours of arriving at the hospital, some PGs within 24 hours prior to delivery were not recorded. We considered that these factors could have influenced the association between hyperglycemia during labor and neonatal hypoglycemia in this study. Maternal prepregnancy obesity (BMI ≥30 kg/m2) is associated with neonatal hypoglycemia [31]. Since Asians, including the Japanese, have impaired insulin secretion, obesity is a rare condition, and Japanese women often develop GDM while underweight [32]. Therefore, maternal obesity may not have been associated with neonatal hypoglycemia in this study. Excessive GWG in GDM mothers was associated with an elevated risk of overall adverse pregnancy outcomes (e.g., LGA, HDP, etc.), but not neonatal hypoglycemia. Although a higher GWG was associated with neonatal hypoglycemia in this study, considering the GWG category, inadequate GWG in the non-neonatal hypoglycemia group was higher than that in the neonatal hypoglycemia group. Thus, it might be better that GWG in GDM mothers is lower than that in non-GDM mothers, considering not only overall adverse pregnancy outcomes, but also neonatal hypoglycemia.
In this study, follow-up of newborn growth up to one month of age showed no significant association with maternal anthropometry, markers of glucose metabolism, or OGTT values at different gestational ages. However, several maternal factors identified in this study significantly increased the risk of hypoglycemia in newborns, suggesting that these factors affect fetal glucose metabolism. Fetal exposure to hyperglycemia in utero affects the development of adipose tissue and pancreatic beta cells and also increases the risk of increased BMI and obesity in older children and adulthood in infants born to mothers with impaired glucose tolerance [3, 33-35]. To date, no reports have examined the association between maternal risk factors for neonatal hypoglycemia identified in this study and the long-term growth of newborns born exposed to them. Further studies of the potential future risk of obesity and metabolic abnormalities in infants born to mothers with the risk factors identified in this study could be important for preventive medicine. The results of this study are significant because they identify new clinical issues in perinatal care.
Our study has some limitations. First, this was a retrospective study with a small sample size, and these kinds of studies have some limitations. However, the sample size was larger than in previous studies. Second, in our hospital, we did not routinely check for PG every neonate. As some neonates born to non-GDM mothers might develop neonatal hypoglycemia, it is unclear whether these findings are only applicable to GDM, or whether it can be extrapolated to all pregnant women. However, this study suggests that abnormalities in glucose metabolism during early pregnancy may affect the short-term prognosis of newborns.
In conclusion, this study showed that HbA1c in the 1st trimester and insulin therapy during pregnancy were good predictors of hypoglycemia in neonates born to GDM mothers, especially when their HbA1c level was 5.2% or more. Our results suggest that abnormalities in glucose metabolism in early pregnancy may affect glucose metabolism in newborns and may be related to their long-term prognosis. We are hopeful that further research will improve the perinatal management of hypoglycemia.
The authors are grateful to all medical staff at Keio University Hospital for excellent patient care. We would like to thank Editage (www.editage.jp) for English language editing.
None.
A.T. collected the data, wrote the manuscript, contributed to the discussion, and reviewed and edited the manuscript; Y.K. collected the data, performed statistical analyses, wrote the manuscript, contributed to the discussion, and reviewed/edited the manuscript; S.I., Y.S., M.H., J.Y., H.I., M.T., and D.O. contributed to the discussion and reviewed/edited the manuscript. All authors approved the final manuscript for publication.
The authors declare that they have no competing interests.
