In Gestational Diabetes, Maternal Ferritin Linked to Lower Infant Developmental Scores at 12 Months

Women with gestational diabetes mellitus carry a distinctive circulating biomarker profile throughout pregnancy, and one component of that profile has now been linked to their infants' early developmental performance. In a prospective case-control study of 240 mother–infant pairs followed from early pregnancy through the children's first birthday, researchers in Chongqing, China, found that women with gestational diabetes showed consistently lower serum retinol and serum vitamin K concentrations alongside markedly higher serum ferritin levels across all three trimesters. When the team traced these trimester-specific biomarkers forward to infant neurodevelopmental performance at 12 months, only ferritin remained independently associated with suboptimal scores in fully adjusted models, while the two fat-soluble vitamins dropped out of statistical significance.

The findings, published in Frontiers in Nutrition on 23 October 2026, are framed by the authors as hypothesis-generating rather than evidence for clinical risk stratification or treatment. Gestational diabetes mellitus is one of the most common metabolic complications of pregnancy, characterized by glucose intolerance and insulin resistance that can disrupt maternal, placental, and fetal metabolic homeostasis. Beyond dysglycemia, the condition coexists with chronic low-grade inflammation, oxidative stress, and broader metabolic dysregulation.

Offspring exposed to the disorder face elevated risks of macrosomia and neonatal hypoglycemia, as well as adverse longer-term metabolic and neurodevelopmental outcomes. Because maternal nutritional and metabolic status can influence fetal neurodevelopment through effects on placental nutrient transfer, substrate availability, oxidative balance, neuronal differentiation, myelination, and synaptogenesis, the research team led by Tingyu Li of Children's Hospital of Chongqing Medical University selected three biomarkers with overlapping but partially distinct biological roles: vitamin A, essential for retinoid signaling, neuronal differentiation, and synaptic development; vitamin K, implicated in sphingolipid metabolism and myelination; and iron, which supports cellular energy metabolism, neurotransmitter synthesis, and myelination in the fetal brain. The study enrolled pregnant women aged 18 to 45 years with singleton pregnancies at Chongqing Yongchuan District Maternal and Child Health Care Hospital between February 2025 and May 2026, excluding those with pre-existing diabetes, chronic hypertension, thyroid disorders, autoimmune diseases, or any condition that could independently affect micronutrient status or fetal neurodevelopment.

All participants underwent universal screening for gestational diabetes at 24 to 28 weeks of gestation using a 75-gram oral glucose tolerance test after a 10-hour overnight fast, diagnosed according to the one-step strategy recommended by the International Association of Diabetes and Pregnancy Study Groups. Fasting venous blood samples were collected at three time points corresponding to the three trimesters, at 8 to 12 weeks, 24 to 28 weeks, and 32 to 36 weeks, and serum aliquots were stored at minus 80 degrees Celsius until analysis. Laboratory personnel were blinded to diabetes status throughout, and all assays were performed in duplicate with intra- and inter-assay coefficients of variation below 8 to 10 percent.

The trimester-specific results were consistent across gestation. Women who met criteria for gestational diabetes had serum retinol concentrations of roughly 0.34 to 0.36 micromoles per liter across pregnancy, compared with 0.45 micromoles per liter in controls, and serum vitamin K concentrations near 0.56 to 0.57 nanograms per milliliter, compared with 0.70 to 0.72 nanograms per milliliter in controls, with all between-group comparisons yielding p-values below 0.001. Serum ferritin moved in the opposite direction: the diabetes group averaged approximately 119 to 123 nanograms per milliliter at each time point, while controls averaged 77 to 83 nanograms per milliliter.

Repeated-measures analyses found no significant group-by-time interactions for any biomarker, meaning the between-group gaps persisted stably across gestation rather than emerging or resolving at particular developmental windows. This stability supported the team's decision to use the arithmetic mean of the three trimester measurements as the summary exposure in the primary regression models. First-trimester dietary data collected with a validated 78-item semi-quantitative food frequency questionnaire provided important context.

Women with gestational diabetes reported significantly lower dietary vitamin A intake, 712.58 versus 788.61 micrograms RAE per day, and lower dietary vitamin K intake, 98.74 versus 105.66 micrograms per day, consistent with their lower circulating concentrations. They also consumed fewer fruits and vegetables, 383.15 versus 423.95 grams per day, and less fiber, while eating more energy, carbohydrate, and fat. Yet dietary iron intake did not differ significantly between groups, 17.94 versus 18.24 milligrams per day, even though serum ferritin levels were substantially higher in the diabetes group.

This discordance suggests that elevated ferritin in gestational diabetes is not explained by iron consumption alone and may instead reflect higher systemic inflammation, altered iron storage or sequestration, or metabolic dysfunction. The inflammatory markers were consistent with this interpretation: maternal C-reactive protein averaged 4.26 versus 2.97 milligrams per liter, and interleukin-6 averaged 3.93 versus 2.18 picograms per milliliter, both elevated in the diabetes group. Offspring neurodevelopment was assessed at 12 months of age by trained developmental pediatricians blinded to maternal diabetes status and biomarker results, using the Chinese standardized version of the Gesell Developmental Schedules.

This instrument evaluates five functional domains, gross motor, fine motor, language, adaptive behavior, and personal-social behavior, with raw scores converted to developmental quotients based on Chinese normative data. Children born to mothers with gestational diabetes scored significantly lower across all five domains. Using the study's prespecified operational threshold of a Total Gesell Score below 85, intended for early identification and surveillance rather than definitive clinical diagnosis, 72.5 percent of diabetes-group offspring met the criterion compared with 31.7 percent of controls.

The authors caution that this inclusive threshold captures borderline scores between 75 and 85, so the prevalence figures should not be equated with clinically diagnosed developmental delay, and that the group means remained within broadly similar interpretive ranges. In fully adjusted multivariable logistic regression, which controlled for maternal age, pre-pregnancy body mass index, years of formal education, gestational age at delivery, infant sex, and continuous maternal HbA1c, higher trimester-averaged serum ferritin was independently associated with increased odds of suboptimal neurodevelopmental performance, with an odds ratio of 1.22 per 10 nanograms per milliliter, a 95 percent confidence interval of 1.004 to 1.34, and a p-value of 0.036. Neither mean serum retinol nor mean serum vitamin K reached statistical significance in the model, with wide confidence intervals suggesting imprecise estimates possibly related to limited sample size.

Among the covariates, older maternal age, earlier gestational age at delivery, and male infant sex were also independently associated with the outcome. Multivariable linear regression for the continuous Total Gesell Score, which additionally included maternal C-reactive protein and interleukin-6, confirmed ferritin as a significant negative predictor with a coefficient of minus 0.05 and a p-value of 0.003, while the overall model explained 28.3 percent of the variance in developmental scores. The team also constructed an exploratory composite Nutritional Risk Score, assigning one point each for serum retinol in the lowest quartile of the cohort distribution, at or below 0.35 micromoles per liter, serum vitamin K in the lowest quartile, at or below 0.57 nanograms per milliliter, and serum ferritin in the highest quartile, at or above 120.43 nanograms per milliliter.

A score of 1, indicating the presence of any one of these three risk categories, was associated with a nearly five-fold higher odds of suboptimal neurodevelopmental performance, with an odds ratio of 4.92 and a 95 percent confidence interval of 1.65 to 14.64. However, the association was not monotonic across higher score categories, with scores of 2 and 3 showing elevated but statistically non-significant odds ratios of 1.43 and 2.66, likely reflecting the small numbers of women in those categories. The authors emphasize that these quartile-defined thresholds are data-driven and population-specific rather than established clinical diagnostic criteria, and that the composite score did not appear to add unique predictive information beyond the continuous biomarker models, which point to ferritin as the primary driver of the association.

Interpreting ferritin requires considerable care, and the researchers confronted this challenge directly. Serum ferritin is simultaneously an iron storage protein and an acute-phase reactant, so elevated concentrations during pregnancy may reflect inflammation, altered iron sequestration, metabolic dysfunction, increased iron stores, or some combination of these processes. Ferritin alone cannot establish iron overload or functional iron excess, and complementary measures such as transferrin saturation and hepcidin were not available in this dataset.

In sensitivity analyses following an adaptation of the BRINDA regression correction approach, the team regressed log-transformed ferritin on log-transformed C-reactive protein and used the residuals to calculate inflammation-adjusted ferritin values. The association with suboptimal neurodevelopmental performance remained statistically significant with these adjusted values, with a logistic odds ratio of 1.34 per 10 nanograms per milliliter and a linear coefficient of minus 0.12, suggesting that inflammation does not fully explain the finding, although residual confounding by unmeasured inflammatory markers cannot be excluded. Exploratory interaction analyses found no statistically significant ferritin-by-diabetes-status interaction, indicating that the ferritin association did not differ significantly by whether the mother had gestational diabetes.

The authors frame their conclusions as hypothesis-generating rather than evidence for clinical risk stratification or treatment. Because the observational design precludes causal inference, ferritin may simply mark underlying metabolic dysfunction not fully captured by the measured covariates, and the neurodevelopmental assessment at 12 months captures only early milestones, longer-term follow-up in independent cohorts is essential before the specific thresholds identified here can be considered for clinical application. Still, the practical implications merit attention.

Ferritin is already measured routinely in many antenatal settings, and the findings raise the question of whether iron management strategies, such as avoiding unnecessary iron supplementation in women with adequate or elevated ferritin, could improve neurodevelopmental outcomes. Future randomized controlled trials testing such approaches, alongside studies incorporating hepcidin, transferrin saturation, and oxidative-stress markers, will be needed to clarify whether the ferritin signal reflects dysregulated iron metabolism, oxidative stress, or a broader metabolic fingerprint of gestational diabetes.

Subject of Research: Maternal serum retinol, vitamin K, and ferritin profiles in gestational diabetes and their association with offspring neurodevelopment at 12 months

Article Title: Maternal serum retinol, vitamin K, and ferritin profiles during pregnancy, gestational diabetes mellitus, and offspring neurodevelopment: a prospective case-control study

Article References: Jiang, C., Pang, Y., Yu, J., Yan, S., Liu, X., & Li, T. (2026). Maternal serum retinol, vitamin K, and ferritin profiles during pregnancy, gestational diabetes mellitus, and offspring neurodevelopment: a prospective case-control study. Frontiers in Nutrition, 13, Article 1967646. https://doi.org/10.3389/fnut.2026.1967646

Image Credits: AI Generated

DOI: 10.3389/fnut.2026.1967646

Keywords: gestational diabetes mellitus, serum ferritin, serum retinol, vitamin K, offspring neurodevelopment, Gesell Developmental Schedules, prospective case-control study, pregnancy biomarkers, inflammation, nutritional risk score, Frontiers in Nutrition, Maternal