Effetto della supplementazione di vitamina D sul controllo glicemico, sulla sensibilità all'insulina e sugli esiti della gravidanza in pazienti con diabete gestazionale sottoposte a terapia dietetica

Introduction

Gestational diabetes mellitus (GDM), defined as glucose intolerance with onset or first recognition during pregnancy [ ], affects approximately 9–25% of pregnancies globally, with notably higher prevalence (17.5%) in Asian populations [ , ]. Beyond immediate obstetric risks, GDM is associated with long-term metabolic consequences for both mother and offspring, including increased risks of type 2 diabetes and cardiovascular diseases [ , ].

Emerging evidence suggests that vitamin D deficiency (< 20 ng/mL) affects 18–84% of pregnant women worldwide and is independently associated with β-cell dysfunction and insulin resistance [ , ] Experimental studies reveal vitamin D’s pleiotropic effects through VDR (vitamin D receptor)-mediated modulation of insulin secretion, inflammatory pathways, and placental angiogenesis [ , ]. However, clinical trials investigating vitamin D supplementation in GDM show conflicting results - while a 2023 meta-analysis ( n = 1,582) demonstrated 32% reduction in insulin requirement [ ], other studies found no significant glycemic benefits [ , ].

Notably, previous trials predominantly used high-dose bolus regimens (≥ 2,000 IU/day) with limited dietary standardization [ ]. Our study uniquely investigates whether low-dose daily supplementation (400 IU) - equivalent to routine prenatal vitamins - provides additive benefits when combined with structured dietary management, addressing a critical gap in clinical practice guidelines.

Materials and methods

### Ethical approval and data source

This retrospective cohort study was approved by the Ethics Committee of Wuxi People’s Hospital, affiliated with Nanjing Medical University (Approval No. 2022-EC-015). Given the use of de-identified data, the requirement for informed consent was waived. All procedures were conducted in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for observational studies and adhered to the ethical principles of the Declaration of Helsinki.

We extracted electronic medical records of 98 consecutive patients diagnosed with gestational diabetes mellitus (GDM) who received dietary management at Wuxi People’s Hospital from June 2022 to June 2024. The inclusion and exclusion criteria were as follows: Inclusion criteria: (1) singleton pregnancy; (2) GDM diagnosis by IADPSG criteria [ ]; (3) documented vitamin D supplementation status. Exclusion criteria: (1) pre-gestational diabetes; (2) incomplete laboratory/pregnancy outcome data; (3) multivitamin use containing vitamin D > 200 IU/day.

### Exposure definition and grouping

Since this is a retrospective cohort study, the Vitamin D and Control groups were defined based on existing medical records and patient self-report, not by randomization. Patients were classified into the following groups: Vitamin D Group ( n = 49): Patients who received ≥ 400 IU/day of vitamin D supplementation for at least 8 weeks during pregnancy. The supplementation status was verified using three sources of data: Electronic prescription records; Nurse-administered medication logs; Patient-reported supplement usage documented in prenatal questionnaires.

Control Group ( n = 49): Patients who did not receive any vitamin D supplementation during pregnancy. This was confirmed through medication reconciliation reports and patient self-declarations.

To minimize misclassification, we cross-validated vitamin D exposure using these three data sources, and any discrepancies in the records were resolved by excluding those patients. Inter-source agreement for vitamin D exposure was high (κ = 0.92).

### Dietary management and covariates

#### Control group

All patients followed standardized dietary management based on FIGO 2015 guidelines. Health education was provided during antenatal check-ups, including GDM-related lectures and encouragement to participate in prenatal classes for expectant mothers.

A personalized dietary plan was developed in collaboration with the nutrition department, which calculated daily caloric intake using the formula: [Height(cm) − 105]×30 + 200 kcal. Macronutrient distribution was as follows: 50–60% carbohydrates (≥ 50% from whole grains); 20% protein; 25–30% fats.

Patients were encouraged to increase daily intake of staple foods (260–315 g), with half of the carbohydrates coming from whole grains (e.g., buckwheat, sweet potatoes).

Patients were also advised on physical activity based on their health conditions, with individualized prescriptions for frequency, intensity, and duration of exercise. Regular follow-up visits were scheduled to monitor progress.

#### Vitamin D group

In addition to the dietary management described above, patients in this group received 400 IU of oral vitamin D drops daily, continuing until delivery.

Covariates collected from patient records included: Baseline 25(OH)D3 levels (if tested within 4 weeks of GDM diagnosis); Seasonal variation in UVB exposure (classified by the month of blood sampling); Physical activity level (assessed via MET-min/week from prenatal exercise logs); Concurrent medications (metformin/insulin use, as documented in treatment plans).

### Observational indicators

#### Primary outcome

The primary outcome was the time to achieve glycemic control, defined as the first occurrence of both fasting plasma glucose < 5.3 mmol/L and 1-hour postprandial glucose < 7.8 mmol/L at two consecutive clinic visits, based on the American Diabetes Association (ADA) 2023 guidelines.

#### Secondary outcomes

Metabolic Parameters: After an overnight fast of at least 8 h, venous blood samples were collected from the median cubital vein. For HbA1c, 2 mL of whole blood was drawn into EDTA anticoagulant tubes, centrifuged at 3,000 rpm for 10 min within 2 h, and analyzed using high-performance liquid chromatography (HPLC; Bio-Rad Variant II Turbo Analyzer). For insulin resistance assessment, 3 mL of blood was collected in serum separator tubes, clotted for 30 min, centrifuged at 3,000 rpm for 15 min, and serum aliquots stored at − 80 °C. Fasting insulin was measured via chemiluminescent immunoassay (Abbott Architect i2000SR), and HOMA-IR was calculated as [fasting insulin (µIU/mL) × fasting glucose (mmol/L)] / 22.5. LDL-C was analyzed enzymatically (Roche Cobas c702 analyzer) using 2 mL of serum processed identically.

Inflammatory and Oxidative Stress Markers: High-sensitivity CRP (hs-CRP) was quantified in lithium heparin plasma tubes: 2 mL of blood was centrifuged immediately at 2,500 rpm for 10 min and analyzed via immunoturbidimetry (Siemens Atellica CH930). Malondialdehyde (MDA) was assessed using EDTA plasma centrifuged at 3,000 rpm for 10 min, stored at − 80 °C, and measured via TBARS assay (Cayman Chemical Kit #10009055).

Pregnancy Outcomes: Pregnancy-related outcomes were assessed, including maternal complications, delivery-related issues, and neonatal adverse events. Maternal complications included pregnancy-induced hypertension, polyhydramnios, premature rupture of membranes, and preterm birth, which were recorded based on clinical diagnosis and obstetric records. Adverse pregnancy outcomes included placental abruption, eclampsia, uterine atony, and postpartum hemorrhage, confirmed through delivery and surgical records. Neonatal outcomes assessed included fetal distress, intrauterine fetal death, neonatal hypoglycemia, and neonatal asphyxia, all recorded immediately after birth and confirmed through neonatal charts.

### Statistical analysis

SPSS version 28.0 was used for all analyses. Propensity Score Matching (PSM) with a 1:1 nearest-neighbor method was applied to balance baseline covariates. The primary outcome was analyzed using Cox regression, adjusted for age, pre-pregnancy BMI, baseline HbA1c, and gestational edema score. Secondary continuous outcomes were assessed with linear mixed models, while categorical outcomes were analyzed using logistic regression. Hypoglycemia incidence was analyzed with modified Poisson regression. Bonferroni correction was applied for inflammatory and oxidative stress markers (α = 0.0083).

### Power analysis

A power analysis was conducted using PASS 2023 software. The target sample size of 98 patients (49 per group) was based on detecting a 30% reduction in time to glycemic control (primary outcome), assuming a median time of 14 days in controls, hazard ratio (HR) = 1.7, α = 0.05, and 80% power. Event probabilities were derived from prior GDM cohort studies at our institution.

### Potential biases

Intervention Compliance: Adherence to vitamin D supplementation was monitored via pharmacy dispensing records and prenatal medication logs. Non-compliant patients ( n = 6) were excluded. This per-protocol approach may introduce bias due to unmeasured time-varying confounders (e.g., changes in diet or physical activity during follow-up), which were not accounted for in the analysis.

Result

### Participant flow and missing data

A total of 132 patients with GDM were initially screened. After applying exclusion criteria (pre-gestational diabetes: n = 10; incomplete data: n = 12; multivitamin use > 200 IU/day: n = 6), 104 patients were eligible for analysis. Six patients (Vitamin D group: 3; Control group: 3) were excluded due to non-compliance (confirmed via medication logs), resulting in 98 patients (49 per group) included in the final analysis (Fig. ).

### Detailed comparisons

After propensity score matching (1:1), 98 patients (49 per group) were included in the analysis. Baseline characteristics were balanced between groups (all P > 0.05), except for significantly higher edema scores in the control group (2.1 ± 0.5 vs. 1.2 ± 0.3, P = 0.001). Both groups had similar vitamin D deficiency at baseline [25(OH)D3: 10.15 vs. 10.05 nmol/L, P = 0.707] (Table ).

### Treatment response

Using multivariable Cox regression adjusted for age, pre-pregnancy BMI, and baseline edema score, the Vitamin D group achieved glycemic control significantly faster than the control group (adjusted HR = 2.30, 95% CI 1.50–3.52, P < 0.001). The mean time to target was 3.67 ± 0.55 days versus 4.11 ± 0.61 days in the control group.

Insulin Dose Reduction: Linear regression adjusted for baseline HbA1c and gestational age showed a significant reduction in insulin requirements in the Vitamin D group (β = -0.11 U/kg/day, 95% CI -0.14 to -0.08, P < 0.001).

Hypoglycemia Risk: Modified Poisson regression adjusted for maternal age and BMI revealed no significant difference in hypoglycemia risk between groups (adjusted RR = 0.45, 95% CI 0.12–1.72, P = 0.240) (Table ).

### Changes in key metabolic and inflammatory markers

Both the vitamin D and control groups showed significant improvements in key metabolic and inflammatory markers after treatment. The vitamin D group exhibited significantly greater reductions in HbA1c, fasting blood glucose, 2hPG, insulin resistance (HOMA-IR), LDL-C, hs-CRP, and MDA compared to the control group (all with P < 0.001). Additionally, the vitamin D group showed a more pronounced decrease in BMI and gestational edema score ( P < 0.001). No significant change was observed in mid-upper arm circumference for either group ( P = 0.712) (Table ).

### Comparison of pregnancy outcomes between the two groups

After adjusting for maternal age, pre-pregnancy BMI, and baseline edema score using multivariable logistic regression, the vitamin D group showed significantly lower risks of pregnancy complications and adverse outcomes compared to the control group (all P < 0.05). Detailed adjusted odds ratios (aORs) and confidence intervals are reported in Tables ( – ).

Discussion

Research indicates that China is a high-incidence country for gestational diabetes mellitus (GDM), with a global prevalence rate as high as 17.5%. It has been reported that vitamin D deficiency in pregnant women is associated with a 39% higher risk of developing GDM compared to women with normal vitamin D levels [ ]. Additionally, the prevalence of GDM has been rising in recent years, with factors such as poor lifestyle choices and advanced maternal age further exacerbating the risk [ ].

Lifestyle modification plays a crucial role in the treatment of GDM, with dietary management serving as a core component of these lifestyle changes. Through a multidisciplinary collaboration model, dietary management improves the control of maternal glucose and lipid metabolism, leading to better maternal pregnancy outcomes and enhanced neonatal immune function [ ]. Studies have shown that combining dietary management with vitamin D supplementation yields optimal intervention results, effectively controlling blood glucose levels and reducing insulin resistance, which in turn enhances pregnancy outcomes [ ]. Furthermore, the combination of dietary management, hypoglycemic medications, and vitamin D supplementation has been found to help partially control blood glucose levels in GDM patients, improve insulin resistance, and reduce adverse pregnancy outcomes such as cesarean sections, postpartum hemorrhage, and negative fetal outcomes [ ].

This study found that the HbA1c of the vitamin D group significantly decreased (from 9.30 to 5.53% vs. 9.25–7.01%, adjusted P < 0.001), as well as fasting blood glucose (from 6.97 to 5.25 mmol/L vs. 6.84 to 5.98 mmol/L, adjusted P = 0.018) and postprandial 2-hour blood glucose (from 10.22 to 7.52 mmol/L vs. 10.13 to 9.21 mmol/L, adjusted P < 0.001), indicating that vitamin D can effectively improve blood glucose control in GDM patients. In addition, the insulin resistance marker HOMA-IR also significantly decreased (from 3.75 to 1.25 vs. 3.82 to 1.80, adjusted P = 0.002), further indicating the potential of vitamin D in improving insulin sensitivity. The mechanism of action is that vitamin D’s improvement of insulin sensitivity is mainly related to its modulation of the insulin signaling pathway through the vitamin D receptor (VDR). Specifically, vitamin D can upregulate the phosphorylation of insulin receptor substrates (IRS), activating the PI3K-Akt pathway, thereby promoting glucose uptake by peripheral tissues [ ]. This is consistent with the decrease in fasting insulin (FINS) observed in this study, indicating that vitamin D helps enhance the effect of insulin. Furthermore, vitamin D also reduces insulin resistance by inhibiting chronic low-grade inflammation and lowering inflammatory factors such as TNF-α and hs-CRP [ ]. Consistent with our findings, the RCT study by Qiu et al. [ ] also found that daily supplementation of 400 IU of vitamin D reduced HbA1c by 0.9% and significantly improved insulin sensitivity. Milajerdi et al. [ ] also showed that vitamin D supplementation significantly reduced hs-CRP and MDA levels, consistent with the reduction in inflammatory markers observed in our study. While the metabolic improvements are evident, the observed BMI reduction in the vitamin D group (from 25.02 to 20.45 kg/m² vs. 25.03 to 23.45 kg/m 2 , adjusted P < 0.001) warrants careful interpretation. Unlike typical weight loss interventions, this change primarily reflected diminished edema, not fat mass alteration Vitamin D may ameliorate edema through dual pathways: Anti-inflammatory Action: By suppressing TNF-α and hs-CRP vitamin D reduces vascular permeability and fluid extravasation, a mechanism previously evidenced in nephrotic syndrome models [ ]. Renal Sodium Handling: Enhanced tubular reabsorption of sodium via upregulation of epithelial sodium channels (ENaC) may counteract pregnancy-induced hypervolumia [ ]. This finding aligns with a trial in preeclampsia where vitamin D supplementation reduced edema incidence by 34% ( P = 0.02) without affecting maternal weight [ ]. Clinicians should therefore interpret BMI changes in GDM management with caution, prioritizing direct body composition metrics over conventional weight-based indices. While these studies showed some effects, the improvements in blood glucose and insulin resistance in our study’s vitamin D group were more substantial, possibly due to the severe baseline vitamin D deficiency in our cohort (10.1 vs. 18.5 nmol/L).

In terms of lipid metabolism, our study found that LDL-C in the vitamin D group significantly decreased (from 3.58 to 2.35 mmol/L vs. 3.60 to 2.85 mmol/L, adjusted P = 0.001), which may be related to vitamin D’s regulation of lipid metabolism and improvement in liver LDL receptor function. Furthermore, MDA levels were significantly reduced in the vitamin D group (from 6.88 to 4.27 mmol/L vs. 6.91 to 5.90 mmol/L, adjusted P < 0.001), suggesting that vitamin D may alleviate oxidative stress, potentially contributing to a reduction in cardiovascular risk. The mechanism is that vitamin D upregulates liver LDL receptor expression, promoting the clearance of LDL, thereby reducing LDL-C levels [ ]. Furthermore, vitamin D’s activation of antioxidant enzymes such as superoxide dismutase (SOD) and glutathione (GSH) may alleviate lipid peroxidation, reducing MDA levels [ ]. These mechanisms indicate that vitamin D plays an important role in improving lipid metabolism, reducing oxidative stress, and providing cardiovascular protection. Additionally, Zeng et al. [ ] also demonstrated that 1000 IU/day of vitamin D significantly reduced HOMA-IR and LDL-C levels, but our study showed that even a lower dose of 400 IU/day of vitamin D could effectively improve lipid metabolism, possibly due to the severe baseline vitamin D deficiency in our study population. Moreover, the role of vitamin D in reducing MDA levels has also been validated in other studies, such as Zhang et al. [ ], where vitamin D supplementation lowered MDA levels, but the effect observed in our study was more significant.

Regarding pregnancy outcomes, the results in the vitamin D group were significantly better than in the control group. The incidence of adverse pregnancy composite events was significantly lower in the vitamin D group (22.45% vs. 57.14%, adjusted P = 0.003). Specifically, the rates of preterm birth (6.12% vs. 12.24%, P = 0.041) and macrosomia (4.08% vs. 10.20%, P = 0.032) were both significantly reduced, suggesting that vitamin D may improve pregnancy outcomes. The mechanism of action may be that vitamin D optimizes placental function, improving pregnancy outcomes. Studies have shown that vitamin D can improve placental perfusion by upregulating placental vascular endothelial growth factor (VEGF), thereby reducing fetal distress [ ]. Additionally, vitamin D may promote the differentiation of regulatory T cells (Treg), inhibit pro-inflammatory Th17 cells, and reduce placental inflammation, thereby lowering the risk of preterm birth [ ]. Similar to the findings of Wang et al. [ ], vitamin D supplementation can significantly reduce cesarean section rates, and in our study, the cesarean section rate in the vitamin D group was significantly lower than in the control group (4.08% vs. 10.20%), which may be related to the combined intervention effect of diet management. Studies [ ] have also reported the potential role of vitamin D in improving pregnancy outcomes, particularly in reducing the incidence of macrosomia and preterm birth, consistent with the findings of our study.

Limitations and future directions

While this study shows that vitamin D supplementation improves glycemic control and pregnancy outcomes in women with GDM, several limitations must be acknowledged. The fixed dose of 400 IU/day may be insufficient for those with severe deficiency, and serum 25(OH)D₃ levels were not monitored to assess adherence or dose-response relationships. Reliance on self-reports and prescription records introduces potential bias. As a single-center retrospective study, findings may be influenced by selection bias and unmeasured confounders such as dietary habits and sunlight exposure. Some outcome measures, including edema and inflammatory markers, were assessed using subjective or single-time-point methods. Long-term maternal and offspring outcomes, as well as the underlying biological mechanisms of vitamin D, were not evaluated. Future research should include multicenter randomized controlled trials with varying vitamin D doses, serial biomarker monitoring, standardized outcome assessments, and long-term follow-up of maternal-offspring health. Omics approaches could further clarify the mechanistic pathways involved.

Conclusion

In conclusion, this study suggests that vitamin D supplementation combined with dietary management significantly improves blood glucose control, insulin sensitivity, lipid metabolism, inflammation, and pregnancy outcomes in GDM patients. The mechanisms may include enhancing insulin sensitivity through activation of the vitamin D receptor, regulating lipid metabolism, alleviating oxidative stress, and improving placental function. Despite limitations such as the dose and sample size, this study provides strong evidence for the application of vitamin D in GDM management, and future research should further explore the optimization of its dose and long-term effects.