Association of diagnostic oral glucose tolerance test values with neonatal hypoglycemia in diet-controlled gestational diabetes mellitus: a retrospective cohort study Authors: Guangxia Zhang, Yunfei Qiu, Jing Su, Juan Du, Shuping Li, Yanchun Fan, Wenjuan Sun, Dongzhen Wang, Lingjie Li, Huifeng Shi Published: 2026 ABSTRACT AbstractBackgroundWhether diagnostic oral glucose tolerance test (OGTT) values are associated with neonatal hypoglycemia (NH) in women with diet-controlled gestational diabetes mellitus (GDM) remains uncertain. We aimed to evaluate the associations of diagnostic OGTT values with neonatal glycemic outcomes and to identify other determinants of NH in this population.MethodsIn this retrospective cohort study, 2,051 singleton pregnancies with diet-controlled GDM were included from a maternal and child health center in Liaocheng, China, between January 2019 and August 2023. Diagnostic fasting blood glucose (FBG), 1-hour post-glucose (1hPG), and 2-hour post-glucose (2hPG) values from the 75-g OGTT were analyzed. Restricted cubic spline (RCS)-based linear and logistic regression models were used to assess possible non-linear associations with neonatal 1-h postnatal blood glucose and NH after adjustments.ResultsNH occurred in 12.0% of neonates (247/2,051). After multivariable adjustment, neither standard regression nor RCS analyses showed significant overall or non-linear associations between diagnostic OGTT values and NH or neonatal 1-h postnatal blood glucose (all P > 0.05). Gestational age at delivery was independently protective, with each additional week associated with a lower risk of NH (aOR = 0.62, 95% CI 0.48–0.80). Cesarean delivery was associated with higher odds of NH than vaginal delivery (aOR = 2.24, 95% CI 1.62–3.10).ConclusionsIn diet-controlled GDM pregnancies, diagnostic OGTT values were not independently associated with NH or neonatal 1-h postnatal blood glucose. These findings suggest that subsequent clinical management and obstetric factors, particularly gestational age at delivery and delivery mode, may be more important than the diagnostic OGTT value itself for neonatal glycemic outcomes.Keywords: Gestational diabetes mellitus, Oral glucose tolerance test, Neonatal hypoglycemia, Neonatal blood glucose, Restricted cubic spline Introduction Hyperglycemia in pregnancy, particularly gestational diabetes mellitus (GDM), is among the most common medical complications of pregnancy and is associated with multiple adverse maternal and neonatal outcomes [1–5]. The prevalence of GDM varies substantially across regions, and its burden has increased in China in recent years [1, 3, 4]. Neonatal hypoglycemia (NH) is one of the most clinically important complications because it may be asymptomatic at presentation but can lead to seizures, neurological injury, and long-term developmental impairment if not recognized promptly [6–13]. Maternal hyperglycemia is a recognized risk factor for NH [14], yet the predictive value of the diagnostic 75-g OGTT for neonatal glycemic status remains uncertain. A single diagnostic OGTT result is attractive as an early risk-stratification tool because it is available in mid-pregnancy, before delivery and neonatal screening [15, 16]. However, the available evidence is inconsistent. Some studies have reported weak positive associations between maternal glucose and NH, whereas others have not identified clinically useful thresholds [16–18]. One important reason for these inconsistent findings is that neonatal glucose homeostasis is influenced not only by diagnostic maternal glucose levels, but also by subsequent glycemic management, fetal maturity, placental function, delivery-related stress, and early feeding practices. This issue may be particularly relevant in women with diet-controlled GDM, in whom post-diagnostic management may attenuate the prognostic effect of the OGTT itself. In addition, few studies have formally evaluated whether the association between OGTT values and neonatal outcomes is non-linear. Therefore, we conducted a retrospective cohort study in women with diet-controlled GDM to examine the association of diagnostic OGTT values with neonatal 1-h postnatal blood glucose and NH, and to explore possible non-linear dose-response relationships using restricted cubic spline (RCS) models. Materials and methods Study design and participants This retrospective cohort study analyzed pregnancies recorded in the electronic health record system of Dongchangfu District Maternal and Child Health Care Hospital, Liaocheng, China, between January 2019 and August 2023. The source database included 2,315 pregnancies complicated by hyperglycemia in pregnancy. After exclusion of 209 women with pregestational diabetes mellitus (PGDM), 2,106 women with gestational diabetes mellitus (GDM) remained. For the present analysis, 18 women who received insulin treatment during pregnancy and 37 women with missing insulin-treatment records were further excluded, yielding a final analytic cohort of 2,051 women with diet-controlled GDM (Fig. 1). Flowchart of participant selection Data collection All study variables were extracted from the hospital electronic health record system. Diagnostic OGTT values and pre-delivery fasting blood glucose were obtained from venous blood samples analyzed in the hospital clinical laboratory using an enzymatic hexokinase assay on a Beckman Coulter AU5800 chemistry analyzer (Beckman Coulter Experimental System [Suzhou] Co., Ltd., Suzhou, China). For fasting venous glucose measurement, 2–4 mL of blood was collected in the fasting state by nurses in hospital. Samples were centrifuged at 3,500 rpm for 5 min to separate serum or plasma and were analyzed promptly. When immediate analysis was not possible, samples were stored and transported according to the laboratory standard operating procedure. The glucose assay was calibrated after instrument maintenance, at the end of each calibration cycle (30 days), and whenever quality-control performance indicated the need for recalibration. Internal quality control was performed daily, and the laboratory participated annually in external quality assessment organized by the Shandong Province Clinical Laboratory Center and the National Health Commission Clinical Laboratory Center. The diagnosis of GDM followed current Chinese guidelines [19]. A 75-g OGTT performed at 24–28 weeks of gestation was used for diagnosis, with fasting blood glucose (FBG), 1-hour post-glucose (1hPG), and 2-hour post-glucose (2hPG) thresholds of 5.1, 10.0, and 8.5 mmol/L, respectively. GDM was diagnosed when at least one of these thresholds was met or exceeded. After diagnosis, all women received standardized lifestyle management according to the five pillars of diabetes care. Patients were counseled to consume six meals per day, emphasize whole grains and high-quality protein, consume at least 300 g of vegetables daily, and limit salt intake to less than 5 g per day. Moderate physical activity, such as walking or swimming, was recommended for 25–30 min after each meal, aiming for at least 1 h of exercise per day. Glycemic control was monitored by weekly outpatient assessment and self-monitoring of fasting and 2-h postprandial blood glucose (target ranges: 3.3–5.3 mmol/L and 4.4–6.7 mmol/L, respectively). Women whose glucose levels remained above target after 5–7 days of lifestyle intervention were treated with insulin and were not retained in the analytic cohort. At our institution, neonates at risk of hypoglycemia were routinely fed as early as possible after birth, with breastfeeding and supplemental formula used according to clinical practice. Neonatal peripheral blood glucose was measured 1 h after birth using an ACCU-CHEK bedside glucose meter (Roche Diagnostics Products [Shanghai] Co., Ltd., Shanghai, China). Instrument maintenance, calibration, and quality control followed the manufacturer’s instructions. At least one internal quality-control procedure was performed on each testing day, and annual external quality assessment was undertaken. NH was defined as a peripheral blood glucose level of less than 2.6 mmol/L at 1 h after birth, which was also the threshold used clinically for early management at our institution [20]. Covariates Covariates considered in the regression models included maternal age, gestational weight gain (GWG), pre-pregnancy body mass index (BMI), gestational age at delivery, parity, delivery mode, hypertensive disorders complicating pregnancy (HDCP), and pre-delivery fasting blood glucose. GWG was calculated as weight at delivery minus pre-pregnancy weight. Pre-pregnancy BMI was calculated as weight in kilograms divided by height in meters squared, using the measurements recorded at the initial antenatal visit. HDCP, including gestational hypertension, pre-eclampsia/eclampsia, chronic hypertension, and chronic hypertension with superimposed pre-eclampsia, was diagnosed by obstetricians according to Chinese clinical practice guidelines [21]. Pre-delivery fasting blood glucose was included as a marker of maternal glycemic control after GDM diagnosis. Statistical analysis Continuous variables were not normally distributed, age, GWG, BMI, gestational age at delivery, gravidity, parity, birth weight, diagnostic OGTT values, and neonatal 1-h postnatal blood glucose were summarized as medians with interquartile ranges (IQRs). Categorical variables were summarized as counts and percentages. Between-group comparisons were performed using the Mann-Whitney U test for continuous variables and the Pearson chi-square test for categorical variables. For descriptive analysis, each OGTT measure was divided into quartiles, and boxplots were used to show the distribution of neonatal 1-h postnatal blood glucose across quartile groups. To assess possible non-linear associations, each OGTT measure (FBG, 1hPG, and 2hPG) was modeled separately using restricted cubic splines (RCS). We fitted multivariable linear regression models with RCS terms for neonatal 1-h postnatal blood glucose as a continuous outcome and multivariable logistic regression models with RCS terms for neonatal hypoglycemia as a binary outcome. Candidate RCS models were compared using the Akaike information criterion (AIC), and the model with the lowest AIC was retained. Overall and non-linear associations were evaluated using Wald tests. We additionally fitted multivariable logistic regression models to estimate adjusted odds ratios (aORs) and 95% confidence intervals (CIs) for NH, and multivariable linear regression models to estimate beta coefficients (β) and 95% CIs for neonatal 1-h postnatal blood glucose. All multivariable models were adjusted for maternal age, pre-pregnancy BMI, GWG, parity, gestational age at delivery, delivery mode, HDCP, and pre-delivery fasting blood glucose. All analyses were performed using R software, version 4.3.2. Two-sided P values < 0.05 were considered statistically significant. Results As shown in Table 1, among the 2,051 eligible mother-neonate pairs included in the final analysis, 247 neonates had hypoglycemia at 1 h after birth, corresponding to an incidence of 12.0%. The median maternal age was 31 years (IQR 27.0–34.0 years). Median OGTT values were 5.20 (5.02, 5.40) mmol/L for FBG, 9.33 (7.85, 10.51) mmol/L for 1hPG, and 7.62 (6.53, 8.73) mmol/L for 2hPG. The median neonatal 1-h postnatal blood glucose level was 3.30 (2.80, 3.90) mmol/L. Table 1Baseline characteristics of participants and neonatesVariablesNeonatal HypoglycemiaOverall, n = 2051P-valueNo, n = 1804Yes, n = 247Age (years), Median (IQR)31.0 (27.0, 34.0)32.0 (28.0, 35.0)31.0 (27.0, 34.0)0.003GWG (kg), Median (IQR)13.5 (10.0, 17.2)12.4 (8.0, 15.7)13.1 (10.0, 17.0)< 0.001BMI (kg/m2), Median (IQR)23.8 (21.5, 27.0)25.0 (22.0, 27.3)24.0 (21.5, 27.0)0.019Gestational age at delivery (weeks), Median (IQR)39.2 (38.4, 40.0)38.6 (38.1, 39.3)39.1 (38.4, 40.0)< 0.001Gravidity, Median (IQR)2.0 (1.0, 3.0)3.0 (2.0, 4.0)2.0 (1.0, 3.0)< 0.001Parity, Median (IQR)2.0 (1.0, 2.0)2.0 (2.0, 2.0)2.0 (1.0, 2.0)< 0.001Delivery Method, n (%)< 0.001 Vaginal934 (52)69 (28)1,003 (49) Cesarean868 (48)178 (72)1,046 (51)HDCP, n (%)0.683 No1,685 (93)229 (93)1,914 (93) Yes119 (6.6)18 (7.3)137 (6.7)FBG (mmol/L), Median (IQR)5.20 (5.04, 5.40)5.20 (4.94, 5.45)5.20 (5.02, 5.40)0.9331hPG (mmol/L), Median (IQR)9.30 (7.81, 10.50)9.50 (8.00, 10.83)9.33 (7.85, 10.51)0.0792hPG (mmol/L), Median (IQR)7.61 (6.53, 8.71)7.73 (6.76, 8.93)7.62 (6.53, 8.73)0.091Neonatal Gender, n (%)0.272 Male1,009 (56)129 (52)1,138 (55) Female795 (44)118 (48)913 (45)Birth Weight (g), Median (IQR)3,350 (3,090, 3,650)3,400 (3,000, 3,720)3,360 (3,080, 3,670)0.3111-h postnatal BG (mmol/L), Median (IQR)3.40 (3.00, 3.90)2.30 (2.10, 2.40)3.30 (2.80, 3.90)< 0.001IQR Interquartile range, GWG Gestational weight gain, BMI Body mass index, GDM Gestational diabetes mellitus, HDCP Hypertensive disorders complicating pregnancy, FBG Fasting blood glucose, 1hPG 1-hour post-glucose, 2hPG 2-hour post-glucose1-h postnatal BG, 1-hour postnatal blood glucoseP values for continuous variables were calculated using the Mann-Whitney U test; P values for categorical variables were calculated using the Pearson chi-square test or Fisher's exact test Baseline characteristics of participants and neonates IQR Interquartile range, GWG Gestational weight gain, BMI Body mass index, GDM Gestational diabetes mellitus, HDCP Hypertensive disorders complicating pregnancy, FBG Fasting blood glucose, 1hPG 1-hour post-glucose, 2hPG 2-hour post-glucose 1-h postnatal BG, 1-hour postnatal blood glucose P values for continuous variables were calculated using the Mann-Whitney U test; P values for categorical variables were calculated using the Pearson chi-square test or Fisher's exact test Figure 2 shows the distribution of neonatal 1-h postnatal blood glucose across quartiles of maternal OGTT values. Across FBG, 1hPG, and 2hPG quartiles, neonatal blood glucose levels showed modest variation, but there was no consistent monotonic decrease or increase across glucose strata. Boxplots of diagnostic OGTT values and neonatal 1-h postnatal blood glucose. Data are from Dongchangfu District Maternal and Child Health Care Hospital, Liaocheng, China, 2019-2023. Distribution of neonatal 1-h postnatal blood glucose according to quartile groups of maternal FBG (A), 1hPG (B), and 2hPG (C) We used RCS-based linear regression to explore potential non-linear associations between maternal OGTT values and neonatal 1-h postnatal blood glucose (Fig. 3). After covariate adjustment, there was no statistically significant overall or nonlinear association for maternal FBG (P-overall = 0.589, P-nonlinear = 0.852) or 1hPG (P-overall = 0.724, P-nonlinear = 0.678) or 2hPG (P-overall = 0.062, P-nonlinear = 0.190). Restricted cubic spline plots for diagnostic OGTT values and neonatal 1-h postnatal blood glucose. Predicted neonatal 1-h postnatal blood glucose values with 95% CIs were estimated for FBG (A), 1hPG (B), and 2hPG (C) using adjusted linear regression models with restricted cubic splines Similarly, RCS-based logistic regression did not show significant overall or non-linear dose-response relationships between diagnostic OGTT values and NH (Fig. 4). For FBG (P-overall = 0.908), 1hPG (P-overall = 0.739), and 2hPG (P-overall = 0.434), no clinically meaningful threshold could be identified. Restricted cubic spline plots for diagnostic OGTT values and neonatal hypoglycemia. Predicted odds with 95% CIs were estimated for FBG (A), 1hPG (B), and 2hPG (C) using adjusted logistic regression models with restricted cubic splines Table 2 summarizes adjusted effect estimates derived from the spline-based models. For continuous OGTT predictors, the reported aORs and beta estimates represent contrasts between the 25th and 75th percentile values of each OGTT measure. After adjustment for maternal and obstetric covariates, diagnostic OGTT measures were not significantly associated with NH or neonatal 1-h postnatal blood glucose. Table 2Adjusted associations of maternal OGTT measures and obstetric factors with neonatal hypoglycemia and 1-hour postnatal blood glucoseNeonatal hypoglycemiaaOR (95% CI)1-h postnatal blood glucoseβ (95% CI)OGTT FBG0.97 (0.86, 1.10)0.01 (-0.01, 0.04) 1hPG1.01 (0.80, 1.26)-0.02 (-0.08, 0.03) 2hPG1.08 (0.86, 1.36)-0.05 (-0.11, 0.00)Obstetric Characteristics Age1.00 (0.79, 1.27)-0.01 (-0.07, 0.04) GWG0.84 (0.69, 1.00)0.01 (-0.04, 0.05) BMI0.95 (0.78, 1.16)-0.02 (-0.07, 0.02) Gestational Week0.62 (0.48, 0.80) **0.17 (0.11, 0.23) ** Parity1.25 (1.00, 1.56) *-0.18 (-0.23, -0.12)Delivery Method Vaginal1.001.00 Cesarean2.24 (1.62, 3.10) **-0.45 (-0.53, -0.39) **HDCP0.96 (0.54, 1.69)0.01 (-0.13, 0.15)Pre-delivery fasting blood glucose1.06 (0.88, 1.29)-0.01 (-0.05, 0.04)OGTT Oral glucose tolerance test, FBG Fasting blood glucose, 1hPG 1-hour post-glucose, 2hPG 2-hour post-glucose, OR Odds ratio, CI Confidence intervalFor OGTT measures, reported estimates represent adjusted contrasts comparing the 25th and 75th percentile values** P < 0.01; * P < 0.05 Adjusted associations of maternal OGTT measures and obstetric factors with neonatal hypoglycemia and 1-hour postnatal blood glucose OGTT Oral glucose tolerance test, FBG Fasting blood glucose, 1hPG 1-hour post-glucose, 2hPG 2-hour post-glucose, OR Odds ratio, CI Confidence interval For OGTT measures, reported estimates represent adjusted contrasts comparing the 25th and 75th percentile values ** P < 0.01; * P < 0.05 In contrast, obstetric characteristics remained important. Each additional week of gestation was associated with a 38% lower odds of NH (aOR 0.62, 95% CI 0.48–0.80) and with higher neonatal blood glucose (β = 0.17, 95% CI 0.11–0.23). Cesarean delivery was associated with higher odds of NH than vaginal delivery (aOR 2.24, 95% CI 1.62–3.10) and with lower neonatal blood glucose (β = -0.45, 95% CI -0.53 to -0.39). Multiparity showed a weaker association with lower neonatal blood glucose, whereas maternal age, GWG, pre-pregnancy BMI, HDCP, and pre-delivery fasting blood glucose were not consistently associated with both neonatal outcomes. Discussion Principal findings In this retrospective cohort of women with diet-controlled GDM, diagnostic OGTT values at 24–28 weeks of gestation were not independently associated with NH or neonatal 1-h postnatal blood glucose after multivariable adjustment. RCS analyses also failed to identify meaningful non-linear associations or clinically useful thresholds. Instead, gestational age at delivery and delivery mode emerged as influential determinants of neonatal glycemic outcomes. Results in the context of what is known Compared with the prospective study by Zhang et al. [17], the incidence of NH in our cohort was lower (12.0% vs. 19.6%). This difference may reflect the more homogeneous population in our study, which excluded PGDM and insulin-treated GDM, as well as standardized antenatal glycemic management and routine early postnatal feeding at our center. These factors may attenuate the relationship between the diagnostic OGTT value itself and neonatal metabolic status. Our findings support the hypothesis that a single diagnostic OGTT has limited prognostic value once maternal glycemia is subsequently managed. Neonatal glucose homeostasis depends on multiple factors beyond the diagnostic glucose value, including later-pregnancy glycemic control, placental function, fetal maturity, intrapartum stress, and early feeding practices [16, 22]. Although HDCP was clinically important and was therefore retained in the adjustment set, it was not independently associated with neonatal hypoglycemia in the final models. Future studies should evaluate whether placental dysfunction or fetal growth patterns modify these relationships. Clinical implications For clinical practice, our results suggest that diagnostic OGTT values alone should not be used to intensify NH surveillance in otherwise well-controlled GDM pregnancies. Greater attention should instead be paid to avoiding non-indicated early delivery and unnecessary cesarean section, because gestational age at delivery and delivery mode showed stronger and more consistent associations with neonatal glucose status than the OGTT values measured at diagnosis. Research implications These findings do not exclude the possibility that other maternal glycemic markers, such as serial fasting glucose, glycated hemoglobin, continuous glucose monitoring metrics, or time-in-range, may better predict NH. Prospective multicenter studies with repeated maternal glucose measurements, detailed treatment adherence data, and standardized postnatal feeding information are needed. Strengths and limitations This study has several strengths. We focused on a clinically relevant and relatively homogeneous cohort by excluding PGDM and insulin-treated GDM. We also used RCS modeling to evaluate both linear and non-linear associations, thereby avoiding the assumption that maternal glucose has a strictly linear relationship with neonatal outcomes. In addition, the study included a relatively large single-center cohort managed within a consistent clinical pathway. Several limitations should be acknowledged. First, the retrospective single-center design introduces the possibility of residual confounding and limits generalizability. Second, we did not have complete longitudinal measures of glycemic control after GDM diagnosis, such as serial glycated hemoglobin or continuous glucose monitoring data, and therefore could not quantify glycemic variability during late pregnancy. Third, detailed data on feeding volume, timing of supplementation, and higher-level neonatal interventions were not fully available. Conclusion In diet-controlled GDM pregnancies, diagnostic OGTT glucose values were not independently associated with NH or neonatal 1-h postnatal blood glucose after multivariable adjustment. The absence of clear linear or non-linear associations suggests that subsequent clinical management and obstetric factors may be more important than the diagnostic OGTT result itself for neonatal glycemic outcomes. Clinical attention should therefore focus on optimizing gestational age at delivery and avoiding non-indicated cesarean section, while future studies should examine longitudinal maternal glycemic measures. Acknowledgements We would like to express our sincere gratitude to the Maternal and Child Health Center of Dongchangfu District, Liaocheng for their invaluable support throughout this study. We appreciate the efforts of all the healthcare professionals and participants who contributed to the successful completion of this research. Abbreviations Gestational diabetes mellitus Neonatal hypoglycemia Oral glucose tolerance test Fasting blood glucose 1-hour post-glucose 2-hour post-glucose Hypertensive disorders complicating pregnancy Gestational weight gain Body mass index Interquartile range Akaike information criterion Odds ratio Adjusted odds ratio Confidence interval Restricted cubic spline Authors’ contributions Guangxia Zhang: Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Supervision, Writing – original draft, Writing – review and editing; Yunfei Qiu: Conceptualization, Formal analysis, Methodology, Software, Visualization, Writing – original draft, Writing – review and editing; Jing Su: Conceptualization, Data curation, Investigation; Juan Du: Conceptualization, Data curation, Investigation; Shuping Li: Conceptualization, Data curation, Investigation; Yanchun Fan: Conceptualization, Data curation, Investigation; Wenjuan Sun: Conceptualization, Data curation, Investigation; Dongzhen Wang: Conceptualization, Methodology, Project administration, Data curation; Lingjie Li: Project administration, Data curation; Huifeng Shi: Conceptualization, Methodology, Project administration, Writing – review and editing, Resources. Funding None. Data availability The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The study was approved by the Medical Ethics Committee of Dongchangfu District Maternal and Child Health Care Hospital, Liaocheng, China (approval No. 20231101). All procedures involving human participants were performed in accordance with the Declaration of Helsinki and relevant local regulations. Because this was a retrospective study based on routinely collected clinical data, the requirement for individual informed consent was waived by the Medical Ethics Committee of Dongchangfu District Maternal and Child Health Care Hospital, Liaocheng, China. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Guangxia Zhang and Yunfei Qiu contributed equally to this work. References Associated Data Data Availability Statement The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.