Gestational diabetes mellitus (GDM) is the most prevalent metabolic disorder during pregnancy, affecting approximately 7%–14% of pregnancies worldwide [1, 2]. In mainland China, a systematic review and meta-analysis of studies that used the IADPSG criteria reported a pooled GDM prevalence of 14.8%. This rate exceeds that reported in many Western populations and may reflect genetic susceptibility, dietary patterns, and increasing maternal age [3]. GDM not only increases the risk for pregnant women, such as gestational hypertension and elevated cesarean section (CS) rates, but also contributes to adverse neonatal outcomes, including unstable blood glucose, macrosomia, and respiratory distress syndrome [4, 5]. Neonatal hypoglycemia (NH) is one of the most common metabolic disturbances in the early neonatal period, and its incidence is markedly higher among infants with established risk factors such as maternal diabetes, growth restriction, or prematurity [6]. In this study, NH was defined as any blood glucose concentration below 2.6 mmol/L within the first 24 hours of life, consistent with widely used operational thresholds for at-risk late preterm and term infants [7]. If not promptly identified and managed, neonatal hypoglycemia can result in acute brain injury that predominantly affects the posterior cerebral regions, as demonstrated by magnetic resonance imaging studies [8, 9]. Long-term follow-up neuroimaging study has revealed that NH is associated with smaller volumes of deep gray matter structures, including the caudate nucleus and thalamus, which persist into mid-childhood [8]. Consequently, timely identification of risk factors is crucial for prevention and early intervention.

Beyond its occurrence, NH exhibits a dose-dependent relationship with subsequent neurodevelopmental outcomes. Data from the Children with Hypoglycemia and Their Later Development (CHYLD) prospective cohort demonstrated that more severe NH (blood glucose 2.0 mmol/L), prolonged episodes, or recurrent events ($\ge$3 episodes) were associated with diminished executive function and visual-motor performance at 4.5 years of age, whereas mild or transient episodes were not associated with these outcomes [10]. In a later follow-up of the same cohort, neuroimaging assessments at 9–10 years of age suggested that NH was associated with structural brain alterations, including smaller caudate volume, which in turn was related to emotional and behavioral outcomes in mid-childhood [11]. Neuroimaging study further corroborates this gradient of risk, with severe NH linked to reduced deep gray matter volume persisting into mid-childhood [8]. Additional systematic reviews suggest that severe or recurrent NH poses greater long-term neurodevelopmental risk than milder forms, although the evidence base remains heterogeneous [12]. Overall, these findings underscore that both the occurrence and severity of hypoglycemic exposure influence subsequent neurodevelopment.

Delivery mode represents an important perinatal factor that may influence neonatal metabolic adaptation, particularly in pregnancies complicated by GDM. CS can mitigate certain obstetric risks, such as dystocia and severe perineal trauma in cases of suspected macrosomia; however, it may also delay the onset of breastfeeding, alter early neonatal stress responses, and increase the likelihood of maternal postoperative complications and impaired early mother–infant contact [13]. In contrast, vaginal delivery (VD) is accompanied by labor-associated hormonal surges and catecholamine release, which may facilitate neonatal glucose homeostasis. These physiological differences have led to ongoing debate regarding whether delivery mode independently contributes to NH risk in infants born to mothers with GDM.

The approach to GDM management may also modify the relationship between maternal diabetes and neonatal outcomes. A recent systematic review and meta-analysis of 18 randomized controlled trials demonstrated that lifestyle intervention, including diet and physical activity, reduced the risk of NH by 27% (relative risk [RR] = 0.73, 95% confidence interval [CI]: 0.54–0.98), with the greatest effect when intervention was initiated before the third-trimester [14]. Furthermore, pharmacologic management with metformin, compared with insulin, has been associated with a 35% lower rate of NH (RR = 0.65, 95% CI: 0.52–0.81), although the underlying mechanisms remain under investigation [15]. These findings suggest that both lifestyle and pharmacologic interventions may attenuate neonatal metabolic risk in pregnancies complicated by GDM.

Several observational studies have examined factors associated with NH in pregnancies complicated by GDM, suggesting that maternal glycemic control, a history of uterine scarring, and delivery mode may influence neonatal glucose instability [16, 17]. In particular, more stringent management of maternal blood glucose has been associated with a lower incidence of NH in some cohorts, implying a potential protective effect of optimized glycemic control [18]. However, these findings are often limited by heterogeneous study populations, variable definitions and timing of hypoglycemia assessment, and incomplete adjustment for key confounders, including maternal body mass index (BMI), diabetes treatment modality (dietary management versus insulin or oral hypoglycemic agents), and the timing of glucose measurements. Previous observational study has suggested that CS may be associated with NH and related early clinical consequences, particularly in the context of adverse maternal or perinatal metabolic conditions [19]. Moreover, the distinction between diet-controlled and pharmacologically treated GDM, both of which may independently influence neonatal glucose homeostasis, has often been overlooked. These limitations underscore the need for more rigorous evaluation of delivery mode as a potential determinant of NH in well-characterized GDM cohorts.

Recent systematic reviews and meta-analyses have further elucidated risk factors for NH in pregnancies complicated by GDM. A 2024 meta-analysis by Wang et al. [20], synthesizing 12 case-control studies encompassing 991 neonates with NH and 4388 controls, found that CS was associated with nearly twofold higher odds of NH (odds ratio [OR] = 1.90, 95% CI: 1.23–2.92). Other identified risk factors included gestational diabetes itself (OR = 1.65), gestational hypertension (OR = 2.79), and small-for-gestational-age status (OR = 2.88). These findings indicate that both maternal characteristics and perinatal factors contribute to neonatal glucose instability.

Despite these robust meta-analytic findings, several knowledge gaps remain. Most previous studies included mixed populations (GDM and non-GDM) or did not comprehensively adjust for key confounders such as uterine scarring, maternal age, gestational age, maternal BMI, and gravidity, factors that are particularly relevant in Asian cohorts with high CS rates. In China, the CS rate remains notably elevated. A recent national analysis of more than 7.6 million births reported an overall rate of 44.5% in 2020, with rates of about 40% even among low-risk term singleton vertex deliveries [21]. Furthermore, few studies have quantified the predictive performance of delivery mode as a standalone risk factor in well-characterized GDM populations. Therefore, in a retrospective study of 201 women with GDM from a tertiary center in China, the present study aimed to: (1) determine whether CS remains independently associated with NH after adjustment for key confounders, and (2) assess its clinical predictive value using sensitivity, specificity, PPV, and NPV.

This retrospective study evaluated the association between delivery mode and NH among 201 pregnancies complicated by GDM. After adjustment for maternal age, gestational age, gravidity, and a history of uterine scarring, CS remained independently associated with an increased risk of NH, with an adjusted OR of 4.48 compared with VD. In addition, analysis of predictive metrics showed that VD was associated with a high NPV (92.2%), which indicates that neonates delivered vaginally were unlikely to develop hypoglycemia, whereas the PPV of CS was relatively modest. These findings suggest that delivery mode may contribute to early neonatal metabolic risk stratification and should be considered in perinatal monitoring strategies for pregnancies complicated by GDM.

Our findings are consistent with previous studies reporting an association between CS and NH. A 2024 meta-analysis by Wang et al. [20] showed that the odds of hypoglycemia were nearly twice as high in infants delivered by CS compared with those delivered vaginally (OR = 1.90, 95% CI: 1.23–2.92). Similarly, a cohort study by Ogunyemi et al. [19] reported that NH in term infants was associated with several obstetric and perinatal factors, including mode of delivery, highlighting the potential role of CS in neonatal glucose instability. Another study of 4602 infants reported that CS was associated with an increased likelihood of requiring glucose infusion (OR = 1.4, 95% CI: 1.1–1.9), with earlier onset of hypoglycemia particularly evident among infants of obese mothers [16]. Our study extends these observations by focusing specifically on pregnancies complicated by GDM, in which we observed a higher OR associated with CS. This stronger association may reflect the unique metabolic environment of GDM: chronic in utero hyperglycemia, persistent neonatal hyperinsulinemia, absence of labor-induced counterregulatory hormonal surges, and delayed initiation of feeding may collectively heighten the risk of hypoglycemia in this subgroup [27, 28, 29]. Previous studies have reported that cesarean delivery is more frequent among neonates with hypoglycemia; however, it may not remain an independent predictor after adjustment for confounding factors, with maternal obesity and glycemic control playing more prominent roles in neonatal glucose regulation [16, 18]. The larger effect size observed in our study population may be attributable to differences in the definition of hypoglycemia, the higher incidence of NH in our population, or variations in confounder adjustment and population characteristics.

The high CS rate observed in our study (48.76% overall) reflects broader trends in China. A 2023 analysis of 7.6 million deliveries across 4359 hospitals reported an overall CS rate of 44.5%, approximately three times the World Health Organization recommended maximum of 15% [21]. Multiple factors have been proposed to contribute to the elevated CS rate in China, including maternal preference, often related to fear of labor pain, provider- and system-level drivers, a high prevalence of repeat CSs, and variation in obstetric practice and capacity across hospitals [30]. In pregnancies complicated by GDM, CS may be indicated more frequently due to concerns about macrosomia, shoulder dystocia, and labor dystocia. Given the association between CS and NH observed in our study, these high CS rates may amplify the population-level burden of neonatal metabolic complications in China.

The increased susceptibility to NH after CS in GDM pregnancies may be explained by several disruptions in perinatal glucose regulation. In GDM, chronic maternal hyperglycemia enhances placental glucose transfer and induces fetal $\beta$-cell hyperplasia, which results in persistent neonatal hyperinsulinemia after birth and impaired activation of gluconeogenesis, glycogenolysis, and ketogenesis [31]. CS may further exacerbate this vulnerability through several mechanisms. Preoperative fasting reduces maternal glucose availability and may limit fetal hepatic glycogen reserves [32]. Delayed skin-to-skin contact and postponed initiation of breastfeeding may reduce early exogenous glucose intake in the immediate postnatal period [29, 33]. From a mechanistic perspective, the absence of labor-related birth signaling hormone surges following CS has been proposed to influence neonatal physiological adaptation [34]. Consistent with this concept, compared with VD, elective CS has been associated with differences in neonatal stress exposure and immediate metabolic adaptation [35]. In addition, differences in early thermal responses between VD and CSs may further contribute to neonatal metabolic stress in the immediate postnatal period [36]. These mechanisms collectively provide a plausible physiological basis for the lower initial glucose concentrations observed in infants delivered by CS in our study population.

Our study has several strengths. It incorporated a comprehensive set of maternal and neonatal characteristics, which enabled a more thorough assessment of risk factors and reduced the likelihood of residual confounding. By adjusting for key variables, including uterine scarring, gestational age, gravidity, and maternal age, the multivariate model provided a more accurate estimate of the independent association between CS and NH. Clinically, these findings have important implications for perinatal care. First, delivery mode should be incorporated into risk stratification algorithms for identification of neonates at elevated risk of hypoglycemia. Second, when CS is indicated in GDM pregnancies, enhanced monitoring protocols should be implemented, including earlier and more frequent glucose measurements. Third, interventions known to reduce NH risk, such as maintaining maternal glucose stability preoperatively, ensuring early skin-to-skin contact, and promoting early breastfeeding, should be prioritized. Individualized delivery decisions should be made through multidisciplinary collaboration, with assessment of maternal glycemic control, estimated fetal weight, and available delivery resources.

CS is an independent risk factor for NH in infants born to mothers with GDM. In clinical practice, clinicians should ensure maternal and fetal safety, consider glycemic control and labor-related risks, select the delivery mode accordingly, and strengthen postpartum neonatal blood glucose monitoring to reduce adverse outcomes associated with hypoglycemia.

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

QL contributed to study conception, data collection, and initial drafting of the manuscript. FW contributed to data analysis, interpretation, and critical revision. SZ contributed to methodology design, statistical analysis, and data curation. BD contributed to overall supervision, final approval of the version to be published, and accountability for all aspects of the work. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

This study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Anhui No.2 Provincial People’s Hospital (Approval No. (R) 2025-107). Because the study used de-identified retrospective data, the requirement for informed consent was waived by the ethics committee.

Not applicable.

This research received no external funding.

The authors declare no conflicts of interest.

Supplementary material associated with this article can be found, in the online version, at https://doi.org/10.31083/CEOG47854.