Fetal movements may hint at motor skills a decade later in children exposed to gestational diabetes

What happens in the womb does not necessarily stay in the womb. A decade-long follow-up of children whose mothers developed gestational diabetes mellitus (GDM) has found that, while most gross motor abilities at around ten years of age look remarkably similar to those of children from uncomplicated pregnancies, three specific motor tasks still set the two groups apart. The findings, published in Frontiers in Clinical Diabetes and Healthcare, suggest that the metabolic environment of a diabetic pregnancy may leave subtle, task-specific fingerprints on motor functioning that remain detectable well into school age.

Gestational diabetes is defined as diabetes first diagnosed or recognized during pregnancy, and it is among the most common metabolic complications of gestation. According to the Developmental Origins of Health and Disease (DOHaD) concept, adverse conditions during sensitive windows of fetal development can trigger long-lasting structural and functional adaptations that shape health across the lifespan. In pregnancies complicated by GDM, the fetus is bathed in an altered metabolic milieu: maternal glucose crosses the placenta freely, potentially driving fetal hyperglycemia and compensatory fetal hyperinsulinemia, while lipid disturbances, altered adipokine and cytokine concentrations, oxidative stress, and changes in placental nutrient transfer add further layers of complexity. These effects need not be visible at birth; they may only surface later in childhood, when neurodevelopmental demands become more intricate.

There is now a growing body of neuroimaging evidence pointing to structural differences in the brains of children exposed to maternal diabetes in utero. Studies have reported reduced cortical thickness in the occipital lobes, the right postcentral gyrus, and the left superior parietal cortex among preadolescent children, along with selective morphological and volumetric changes in the hippocampus, microstructural white matter abnormalities in infants, and reduced cortical excitability and neuroplasticity in children aged 11 to 13. Because successful motor performance requires the coordinated interaction of motor planning and execution, sensory feedback, visuomotor processing, and postural control, gross motor abilities serve as a sensitive neurodevelopmental indicator. Motor and cognitive development are closely intertwined, especially for complex tasks that cannot be completed without typical cognitive functioning, making motor assessment a practical window onto whether an altered fetal metabolic environment is associated with long-term developmental outcomes.

The new study is a 10-year follow-up of a prospectively established cohort in which fetal neurobehavior had previously been assessed. The original sample comprised pregnant women with singleton pregnancies between 32 and 38 weeks of gestation: 82 women diagnosed with GDM and 82 women with uncomplicated pregnancies and normal oral glucose tolerance test results. GDM was diagnosed using a 75-g OGTT according to the HAPO criteria, with fasting plasma glucose of at least 5.1 mmol/L, 1-hour plasma glucose of at least 10.0 mmol/L, and 2-hour plasma glucose of at least 8.5 mmol/L. Critically, all women with GDM were managed with dietary treatment only, and none received insulin therapy, meaning the cohort represents the milder end of the disease spectrum. Fetal neurobehavior was evaluated during the third trimester using the modified Kurjak Antenatal Neurodevelopmental Test (KANET) with four-dimensional ultrasound, which enables standardized evaluation of both the quantity and quality of fetal movements, from isolated head anteflexion and eye blinking to hand-to-face movements and the overall impression of general movements. The fetal assessor was blinded to GDM exposure status.

At the 10-year follow-up, 55 of the original 82 GDM-exposed families and 54 of the 82 unexposed families could be contacted, corresponding to contact rates of 67.1 and 65.9 percent. After exclusions for genetic developmental difficulties, perinatal or postnatal complications, or acquired impairments, the final sample comprised 50 children in each group. The GDM-exposed group included 36 girls and 14 boys, while the unexposed group included 20 girls and 30 boys, an imbalance that reflects the composition of the original cohort rather than sex-based selection. The children’s mean age was 10 years, with a range from 9 years and 3 months to 11 years and 1 month, deliberately before pubertal maturation typically introduces substantial biological and psychomotor changes that would complicate interpretation. Importantly, the two groups did not differ in anthropometric characteristics, including body weight, height, and body fat percentage.

Gross motor functioning was assessed across six domains: coordination, agility, flexibility, explosive strength, muscular endurance, and balance. Coordination was measured with the Backwards Obstacle Course and Ball Rolling with the Non-Dominant Hand tests, agility with the Side Steps and Shuttle Run Transfer tests, flexibility with the Straddle Forward Bend and Bench Forward Bend tests, explosive strength with the Standing Long Jump and 20-meter Sprint, and muscular endurance with the Sit-Up and Squat tests, all drawn from the CROFIT norm battery. Balance was assessed separately using quiet standing with eyes closed on a specially designed unstable platform, with the Gyko inertial sensor system recording platform displacement path length, velocity, mean distance, mean frequency, and root mean square distance. The kinesiologist conducting the childhood assessments was blinded to prenatal GDM exposure status.

Because several motor outcomes deviated from normality and subgroup sizes were relatively small, the researchers employed an Aligned Rank Transform (ART) factorial ANOVA to examine the effects of group, sex, and the group-by-sex interaction, with effect sizes expressed as partial eta squared. Benjamini-Hochberg false discovery rate (FDR) correction was applied to the p-values for the main effect of group to account for multiple comparisons, performed separately for the motor performance outcomes and the five balance parameters derived from the same quiet-standing assessment. A post hoc power analysis showed that the final sample of 100 participants provided a statistical power of 0.74 for detecting a medium effect size, below the conventional threshold of 0.80, so non-significant findings were interpreted cautiously.

The headline result is one of reassurance with nuance. Gross motor functioning was largely comparable between the groups, with no significant main effect of group for sprint, standing long jump, lateral steps, shuttle run, sit-ups, backward obstacle course, bench forward bend, or any of the balance parameters. However, three specific motor tasks remained statistically significant after FDR correction: children in the GDM-exposed group performed poorer in squats (FDR-adjusted p = 0.027, partial eta squared = 0.079) and ball rolling with the non-dominant hand (FDR-adjusted p = 0.001, partial eta squared = 0.152), while they performed better in the straddle forward bend (FDR-adjusted p = 0.027, partial eta squared = 0.072). The magnitude of the group effects was moderate for squats and the straddle forward bend and large for ball rolling with the non-dominant hand. Sex showed significant main effects for several motor outcomes, and significant group-by-sex interactions were observed for sprint, shuttle run, and ball rolling, highlighting the importance of considering sex when interpreting motor performance in this age group.

The task-specific pattern is what makes these findings scientifically intriguing. No difference was found in coordination during the backward obstacle course, yet the GDM group achieved a poorer result when rolling a ball with the non-dominant hand. Rolling a ball with the non-dominant hand requires coordination of the upper limbs, continuous adjustment of movement direction and intensity, spatial orientation, visual tracking, and control of the non-dominant side of the body, placing greater demands on visuomotor integration and bimanual organization of movement. This interpretation is partly consistent with earlier work by Bolaños and colleagues, who reported poorer graphic, spatial, and bimanual abilities and a higher frequency of minor neurological signs among school-aged children born from GDM pregnancies. Similarly, in muscular endurance, no difference emerged in sit-ups, but the GDM group completed significantly fewer squats within one minute, a task involving repeated dynamic activation of the lower-limb and trunk muscles, postural control, and effective temporal and spatial organization of a repeated movement pattern. The better straddle forward bend performance was unexpected, particularly because no significant difference was found in the bench forward bend; the two protocols are performed from different starting positions and may not reflect the same components of flexibility equally. One potential, though still insufficiently studied, biological factor is the hormone relaxin, which remodels connective tissue and collagen. Numerically higher plasma relaxin concentrations have been observed in women with GDM, and higher relaxin-2 levels correlated with C-peptide concentrations in early GDM, but whether altered maternal relaxin during pregnancy can affect fetal connective and musculoskeletal development is unknown, so this remains a biological hypothesis rather than an explanation of the results.

The authors emphasize that these findings do not indicate generalized gross motor impairment but suggest that prenatal exposure to the metabolic environment associated with GDM may be related to subtle, task-specific differences in motor functioning that remain detectable during school age. The encouraging absence of differences in most abilities does not support the assumption of a general motor deficit later in childhood. Still, the study has limitations: the sex distribution differed between groups, the GDM-exposed group included only diet-treated pregnancies so findings may not generalize to more severe or insulin-treated GDM, and the long interval between fetal assessment and follow-up without intermediate developmental assessments means individual trajectories could not be reconstructed. Confounders such as socioeconomic status, maternal pre-pregnancy BMI, gestational weight gain, and maternal education were not included in the analyses. The researchers suggest that more detailed monitoring of motor functioning may be useful for children born from GDM pregnancies who also show coordination difficulties, reduced motor competence, or lower participation in physical activities, and that regular, developmentally appropriate, structured activities focusing on coordination, balance, movement control, and muscular endurance should be provided to those with difficulties. Further longitudinal research with larger samples and detailed information on maternal metabolic factors, physical activity, and neurodevelopmental functioning is needed to determine the developmental stability and clinical relevance of these outcomes.

Subject of Research: A 10-year follow-up study finds mostly comparable gross motor function at age ten in children prenatally exposed to diet-treated gestational diabetes, with subtle task-specific differences in squats, non-dominant hand ball rolling, and straddle forward bend flexibility.

Article Title: Fetal behavior as a window to later gross motor function: a 10-year follow-up of children prenatally exposed to gestational diabetes mellitus

Article References: Vasilj, O., Marinić, D., Trošt, T., & Žeželj, Z. (2026). Fetal behavior as a window to later gross motor function: a 10-year follow-up of children prenatally exposed to gestational diabetes mellitus. Frontiers in Clinical Diabetes and Healthcare, 7, Article 1938676. https://doi.org/10.3389/fcdhc.2026.1938676

Image Credits: AI Generated

DOI: 10.3389/fcdhc.2026.1938676

Keywords: gestational diabetes, fetal behavior, gross motor development, neurodevelopment, DOHaD, pediatric health, longitudinal study, kinesiology, Fetal, behavior, window, later