This study was a single-center research in northern China, latitude ∼38 °N. A total of 208 children diagnosed with Autism Spectrum Disorder (ASD) were selected as the ASD group, all of whom received treatment in the Psychological Behavior Department of our hospital from January 2021 to December 2023. The study group included 208 children aged between 1 and 5 years, comprising 112 boys and 96 girls. Simultaneously, the control group was established consisting of another set of 208 children aged between 1 and 5 years from the Children's Health Care Department of our hospital. These control subjects were selected through routine physical examinations and also included an equal distribution of genders: 112 boys and 96 girls. Importantly, members of the control group exhibited no symptoms indicative of ASD following preliminary diagnosis and assessment by qualified medical professionals. This study received approval from the Ethics Management Committee overseeing medical clinical trials at our hospital (NO: 202136), with informed consent obtained from the guardians of all participants.

The diagnosis of ASD in participating children adhered to criteria outlined in the Diagnostic and Statistical Manual of Mental Disorders—Fifth Edition (DSM-5) (11) published by the American Psychiatric Association, as well as utilizing the revised Autism Diagnostic Observation Schedule (12).

Inclusion criteria: 1. compliance with established diagnostic criteria for ASD; 2. voluntary participation in this study along with cooperation in completing relevant diagnostic assessments and scale scoring; 3. children diagnosed as ASD for the first time; 4. no history of vitamin D supplementation within three months prior to enrollment.

Exclusion criteria: Participants exhibiting other conditions that may confound this study will be excluded, including: 1. children whose main diagnosis was not autism; 2. children without perfect systematic examination; 3. children who cannot cooperate with the study.

Laboratory tests for 25(OH)D were performed by personnel from the pharmacological laboratory of the Department of Pharmacy. Two ml venous blood was collected, centrifuged for 6 min to obtain serum, and stored frozen in a refrigerator at −20 °C for subsequent test. The stability test confirmed that the samples should be frozen (−20 °C)for no more than 1 month. The half-life of vitamin D was 24 h (13), while the half-life of 25(OH)D was 3 weeks, and the half-life of 1,25-(OH) 2-D was 4 h (14). Due to the long half-life of 25(OH)D, the serum concentrations of 25-hydroxyvitamin D₂ [25(OH)D₂] and 25-hydroxyvitamin D₃ [25(OH)D₃] were used to represent the serum concentrations of vitamin D₂ and vitamin D₃ in the actual detection process respectively. The contents of 25(OH)D₂ and 25(OH)D₃ in serum were detected by high-performance liquid chromatography tandem mass spectrometry (HPLC-MS/MS) method.

Samples were run on a Jasper™ HPLC system with a Triple Quad™ 4500MD mass spectrometer from AB SCIEX (Framingham, MA USA). Analyst 1.6.3 and MultiQuant™ MD 3.0.2 software were used for instrument control, data acquisition, and data processing. Chromatographic separation was performed on a Kinetex® EVO C18, 2.600 µm, 30 × 2.100 mm HPLC column (Phenomenex, Torrance, CA, USA) using a binary gradient elution 0–0.500 min, 80%B; 0.500–2.000 min 98%B; 2.000–5.000 min 98%B; 5.000–6.000 min 80%B; 6.000–7.000 min. The mobile phase (A) and (B) were from Shandong Yingsheng Biotechnology Co., Ltd and the flow rate was 0.550 ml/min. The injection volume was 5 µl. The column oven was kept at 45 °C. The autosampler temperature was maintained at 5 °C.

Mass spectrum acquisition was performed in positive electrospray ionization mode with multiple-reaction monitoring (MRM). Mass parameters were set: temperature, 600 °C; curtain gas, 35 psi; ion spray voltage, 5,500 V; ion source gas 1, 50 psi; and ion source gas 2, 55 psi. Dwell time was set at 30 ms for each compound. The declustering potential, entrance potential, and collision energy were optimized at 92 V, 10 V, and 28 eV, respectively. LC-MS/MS data were analyzed using Analyst 1.6.3 software.

The autism behavior checklist (ABC) (15) contained 57 items, which encompassed five factors, such as sensory (S), relating (R), body and object use (B), language (L), and social and self-help skills (V) (16). In 1989, Professor Yang of Beijing Medical University introduced and revised the score, which were used to screen autistic children. Besides, Childhood Autism Rating Scale (CARS) listed 15 items, and each was graded from 1 to 4 (17). ASD severity was evaluated using ABC and CARS. Higher scores of ABC and CARS scores represented more serious autistic symptoms. Intellectual and behavioral development in ASD children was assessed via the revised Gesell Developmental Schedule (GDS). The evaluation results were presented in the form of a developmental quotient (DQ). GDS covered five sectors, including adaptive behavior, gross motor, fine motor, language, and personal-social behavior (18). The higher development quotient (DQ) scores, the closer the developmental level was to or exceeds that of normal children. During the assessment of ABC, CARS and GDS for children with autism, it was found that many children could not socialize normally, resulting in that many children did not complete all assessment of ABC, CARS and GDS. Many patients had selectively completed one, two or three assessments according to individual differences. In this study, the specific number of participants in each project were as follows: 74 cases were scored with the ABC; 74 cases were scored with the CARS; 60 cases were scored with the GDS. These cases were strictly screened according to the above criteria by two workers. The two doctors took up their posts after unified training. Besides, they were specially responsible for the evaluation of children and had rich experience.

SPSS 21.0 software was used to analyze the data. Normally distributed data were expressed as mean ± standard deviation (SD). Numerical variables without normal distribution were shown as mean ± SD, as well as median values and interquartile range (IQR). The difference of measurement data was detected by t test. If the data were non-normal distribution, Mann–Whitney U test was used. Pearson correlation test was used to test the correlation of parameters. All analyses in this study were bilateral tests. Differences were deemed statistically significant when p < 0.050.

This research project was derived from Medical Science Research Project of Hebei (NO: 20220772). The Health Commission of Hebei Province had approved this project. Guidance on policies and directional resources were provided. The corresponding authors were responsible for all aspects of the study. The final version was approved by all authors.

The serum levels of 25(OH)D₂ and 25(OH)D₃ were assessed in children with ASD group compared to healthy controls. Notably, the level of 25(OH)D₂ in the blood of children within the ASD group was significantly lower than that observed in the control group [(0.913 ± 0.419) vs. (6.841 ± 5.230) nmol/L, P = 0.001]. Similarly, the level of 25(OH)D₃ in the ASD group was significantly reduced compared to controls [(26.862 ± 8.200) vs. (31.953 ± 9.400) nmol/L, P = 0.015] (Table 1, Figure 1).

Pearson linear correlation analysis was conducted to examine the relationship between serum levels of 25(OH)D₂, 25(OH)D₃ and ABC scores among a subset of 74 ASD patients (34 boys and 40 girls). No statistically significant correlation was found between the level of 25(OH)D₂ and ABC scores (r = −0.027, P = 0.872). In contrast, the lower 25(OH)D₃ levels were associated with higher ABC scores. (r = −0.35, P = 0.034).

Pearson linear correlation analyses were performed on serum levels of both 25(OH)D₂, 25(OH)D₃ concerning CARS scores among another set of ASD patients (44 boys and 30 girls). The results indicated no significant association between levels of serum 25(OH)D₂ and CARS scores (r = 0.200, P = 0.236). However, a notable correlation was identified for the lower levels of serum 25(OH)D₃ with higher CARS scores (r = −0.453, P = 0.048) demonstrating statistical significance.

Additionally, an analysis involving pearson linear correlation examined serum concentrations of both forms of 25(OH)D₂, 25(OH)D₃ alongside GDS assessments from a sample size comprising 60 patients with ASD (34 boys and 26 girls). As presented in Table 2, there existed a significant negative correlation that the higher level of 25(OH)D₂ was significantly correlated with lower adaptive behavior and personal-social of GDS (P < 0.050). Such significance was not noted regarding gross motor, fine motor or language dimensions (P > 0.050). Conversely, 25(OH)D₃ demonstrated positive correlations that the higher level of 25(OH)D₃ was significantly correlated with higher adaptive behavior, fine motor and personal-social of GDS (P < 0.050). But it showed no associations with gross motor or language dimensions in Table 3 (P > 0.050).