We retrospectively analyzed clinical data collected from 159 school-aged children (aged 6–12 years) during health examinations at the Hainan Hospital of PLA General Hospital and Lingshui Li Autonomous County People’s Hospital from April 2019 to April 2022. Baseline data were systematically collected using a standardized data collection form to ensure consistency and accuracy, which included demographic information such as age, gender, and residential location obtained through parental interviews and hospital records. Anthropometric measurements, including height and weight, were recorded using calibrated equipment, and BMI was calculated using the formula: BMI = weight (kg) /height² (m²). Serum samples were collected to measure TC, TG, Ca, ALP, FBG and HbA1c in accordance with standard laboratory protocols. Data pre-processing involved detecting outliers using the interquartile range (IQR) method and handling missing values via multiple imputation techniques to ensure data integrity. To facilitate comparability between datasets from the two hospitals, Z-score normalization was performed on the biochemical parameters. The American Academy of Pediatrics’ criteria for vitamin D deficiency were applied: <15 ng/mL as deficiency, 15–20 ng/mL as insufficiency, and ≥20 ng/mL as sufficiency. In this study, participants were categorized into two groups based on their 25(OH)D₃ serum levels: the deficiency group (n = 56), including both deficient (<15 ng/mL) and insufficient (15–20 ng/mL) levels; the non-deficiency group (n = 103), with sufficient levels (≥20 ng/mL).

The inclusion criteria were as follows: (1) All participants were children undergoing routine health examinations. (2) Age range of 6 to 12 years. (3) Availability of complete clinical data for each participant.

The exclusion criteria were as follows: (1) Physical growth abnormalities or deformities. (2) A history of gastrointestinal, respiratory, or other system-related illnesses and recent medication use within the past 3 months. (3) Immunological disorders or uncontrolled infections. (4) Impaired heart, liver, or kidney functions.

Ethics approval and consent to participate: From 2019 until late 2022, clinical data collected from 159 school-aged children during health examinations and outpatient visits at the Hainan Hospital of PLA General Hospital and Lingshui Li Autonomous County People’s Hospital. The Ethics Committee of Hainan Hospital of PLA General Hospital and the Ethics Committee of Lingshui Li Autonomous County People’s Hospital was responsible for supervising and approving the course of the entire study.

A total of 5 mL of venous blood was obtained from each child through standard venipuncture techniques, ensuring aseptic conditions. Blood samples were processed within 2 h post-collection to maintain the integrity of the biochemical analyses.

TC and TG levels were measured using a ADVIA 2400 automated biochemical analyzer (Siemens AG, Germany). Serum was separated from whole blood via centrifugation at 3000 rpm for 10 min at room temperature. Serum samples are mixed with the corresponding reagents as per the manufacturer’s instructions, and the resulting TC and TG levels are assessed colorimetrically at 500 nm. All reagent kits are sourced from Roche (Shanghai, China).

FBG levels were assessed using the 7060 automated biochemical analyzer (Hitachi, Japan). Following serum preparation as described, the glucose oxidase method was utilized, wherein glucose is oxidized to gluconic acid and hydrogen peroxide. The resultant hydrogen peroxide is colorimetrically measured at 505 nm. The reagent kits are from Roche (Shanghai, China).

The levels of HbA1c were analyzed using the RT6000 microplate reader (Rayto, China). Whole blood was mixed with a buffer, incubated at room temperature for 15 min, and then centrifuged at 3,000 rpm. The supernatant was subsequently injected into the HPLC system, with measurements taken at an absorbance of 415 nm. Reagent kits were supplied by Roche (Shanghai, China).

Calcium and ALP levels were measured using the Hitachi 7060 automated biochemical analyzer. After centrifugation of serum at 4°C, serum samples were mixed with the reagents for calcium and ALP assays in accordance with Roche (Shanghai, China) instructions. The calcium levels were measured at 650 nm, while the ALP levels were assessed at 405 nm.

The measurement of 25(OH)D₃ levels was performed using an enzyme-linked immunosorbent assay (ELISA) by Roche (Shanghai, China). Serum samples were diluted in a buffer, followed by the addition of specific antibodies against 25(OH)D₃. After incubation and washing, a substrate solution was added to produce a color change proportional to the 25(OH)D₃ concentration in the sample. Absorbance was measured at 450 nm using the 680 microplate reader (Bio-Rad, United States).

All children underwent a radiographic examination of the left wrist in the hospital after admission using a Shimadzu X-ray machine. During the radiographic examination, the palm and fingers were placed facing downwards on the detector. The children were instructed by the physician to extend their forearms and hands until the forearm axis aligned with the middle finger axis, with their fingers naturally spread apart. The central axis of the X-ray beam was aligned with the third metacarpophalangeal joint gap, and a perpendicular projection was taken, with the focal-film distance set at 90 cm. Furthermore, the BMI was calculated following the collection of admission statistics for the children, using the formula BMI = body weight (kg) /height² (m²).

Data analysis was performed using SPSS 22.0. Continuous variables were assessed for normality using the Kolmogorov–Smirnov test. Independent sample t-tests were employed for normally distributed variables, while non-normally distributed data were analyzed using the Mann–Whitney U test. Categorical data were compared using the χ² test. Spearman correlation analysis was performed to explore associations between 25(OH)D₃ levels and BMI, TC, TG, Ca, ALP, and bone age. Multivariate logistic regression analysis was conducted to assess the independent predictors of 25(OH)D₃ deficiency. ROC curve analysis was used to evaluate the predictive value of BMI, TC, TG, Ca, ALP, and bone age in diagnosing vitamin D deficiency. Statistical significance was set at p < 0.05.

To explore the correlation between vitamin D deficiency and the prevalence of obesity in children, we conducted a retrospective analysis of the clinical data from 159 school-age children who underwent physical examinations (Figure 1). Among the participants, 109 were male and 50 were female, with ages ranging from 6 to 12 years and an average age of 9.27 ± 2.10 years. No significant differences were observed in gender, age, height (Figure 2A), FBG, and HbA1c between the two cohorts (p > 0.05). However, the group with 25(OH)D₃ deficiency showed significantly higher levels of weight, BMI, TC, TG, and ALP compared to the non-deficient group (Figures 2B–F). In contrast, Ca levels and bone age were significantly lower in the 25(OH)D₃ deficiency group (Figures 2G,H). The general characteristics of these patients are detailed in Table 1.

In the present investigation, we collected anteroposterior radiographs of the left hand from a cohort of school-age children for bone age evaluation. Furthermore, we have featured a subset of exemplary radiographic findings. Figure 3A presents a standard anteroposterior X-ray image of the left wrist from a 9-year-old female participant in the non-25(OH)D₃ group. The radiographic R-series indicates a bone age advancement of 9 months relative to chronological age, while the C-series demonstrates a normal bone age alignment. Figure 3B likewise represents an anteroposterior radiograph of the left wrist from a 9-year-old male, the R-series suggests a bone age 5 months older than the actual age, and the C-series indicates a 1-year and 2-month difference. Figure 3C shows a radiograph from a 9-year-old male exhibiting 25(OH)D₃ insufficiency. The R series indicates a bone age 2 years younger than his actual age, while the C series suggests a bone age 1 year and 10 months older. Similarly, we found that school-aged children within the 25(OH)D₃ deficiency cohort exhibit lower R-series and C-series ages than their peers (Figure 3D). In summary, our research findings indicate that a deficiency in 25(OH)D₃ during the school-age period is a significant risk factor for delayed growth and development.

We further constructed the receiver operating characteristic (ROC) curve to assess whether BMI, TC, TG, Ca, ALP, and bone age could be utilized for forecasting 25(OH)D₃ deficiency in school-aged children. The AUC for each parameter was BMI (AUC: 0.879), TC (AUC: 0.785), TG (AUC: 0.872), Ca (AUC: 0.839), ALP (AUC: 0.688), and bone age (AUC: 0.809), respectively, all of which presented a significance level of p < 0.05 (Table 2). The results confirmed that a vast majority of clinical Indicators have a significant diagnostic value (Figure 4).

To further analyze whether there is a direct correlation between 25(OH)D₃ levels and BMI, BMI, TC, TG, Ca, ALP, and bone age. We selected 56 children from the 25(OH)D₃ deficiency cohort and used Spearman’s correlation analysis to assess whether there is an association between their serum 25(OH)D₃ levels and BMI, TC, TG, Ca, ALP, and bone age. The outcomes illustrated that there was a negative correlation between 25(OH)D₃ levels in the children’s serum and BMI (R² = −0.6234, p < 0.0001), TC (R² = −0.5998, p < 0.0001), TG (R² = −0.6125, p < 0.0001), and ALP (R² = −0.6019, p < 0.0001) (Figures 5A–D), whereas a positive correlation was observed with Ca (R² = 0.5920, p < 0.0001) and bone age (R² = 0.6729, p < 0.0001) (Figures 5E–F). These correlations were statistically significant with p < 0.05, as presented in Table 3.

To further explore the relationship between 25(OH)D₃ levels and BMI, TC, TG, Ca, ALP, and bone age, we performed a logistic regression analysis. The dependent variable was the presence of 25(OH)D₃ deficiency, while BMI, TC, TG, Ca, ALP, and bone age were included as independent variables in the model. The logistic regression analysis revealed that Ca, ALP levels and bone age were significant independent predictors of 25(OH)D₃ deficiency, with odds ratios (ORs) greater than 1 (p < 0.05), indicating an increased risk of deficiency (Figure 6A). To further substantiate our findings, we developed a ROC curve using multiple indicators (BMI, TC, TG, Ca, ALP, and bone age), which demonstrated the strong predictive value of these factors for identifying 25(OH)D₃ deficiency (Figure 6B). These findings further support the significant correlations observed in the Spearman analysis, demonstrating the strength of association between these metabolic and bone-related markers and 25(OH)D₃ deficiency.

Based on our findings, vitamin D deficiency is closely linked to obesity in school-age children. We further conducted a subgroup analysis of children with 25(OH)D₃ deficiency stratified by their BMI levels. Our analysis revealed that 25(OH)D₃ deficient children with higher BMI had significantly elevated levels of TC, ALP, and advanced bone age, while Ca levels were relatively lower (Figures 7A–F). These findings highlight a stronger association between obesity and specific metabolic and bone-related markers in children with 25(OH)D₃ deficiency, providing further insight into the complex interplay between 25(OH)D₃ status and childhood obesity.