This was a retrospective cohort study conducted at the Guangxi Medical University First Affiliated Hospital Reproductive Center. We enrolled 188 infertile patients who underwent routine IVF and ICSI treatments between June 2020 and July 2021. Patients were categorized into two groups based on clinical pregnancy outcomes: those who achieved successful pregnancy and those who did not. This study was approved by the Institutional Review Board of Guangxi Medical University First Affiliated Hospital (project number: IRB2020-1234, date of approval: 1 May 2020). All methods were performed in accordance with the Declarations of Helsinki, and written informed consent was provided by all patients before the study.

The inclusion criteria were as follows: patients who underwent fresh embryo transfers, complete follow-up records, no history of genetic diseases in either partner, no uterine abnormalities or concurrent adenomyosis in female patients, all participants were part of a controlled ovulation induction program, and none had taken or injected vitamin D-related preparations within the previous 6 months.

The exclusion criteria were as follows: cancellation of the embryo transfer for any reason; recipients of a frozen embryo transfer; patients with incomplete follow-up; and the presence of systemic diseases, such as diabetes, kidney disease, hypertension, or immune system disorders.

Comprehensive data were collected, including baseline demographic and clinical characteristics [age, body mass index (BMI), duration of infertility, type of infertility], biochemical markers (AMH, AFC, and basal hormone levels including FSH, LH, E2, P, PRL, T, DHEA), treatment details (COH protocol, total Gn dosage, duration of Gn administration), and outcomes (endometrial thickness on the day of human chorionic gonadotropin (HCG) administration, E2 levels on HCG day, number of oocytes retrieved, number of embryos obtained, clinical pregnancy outcome).

Vitamin D levels were measured using a chemiluminescence immunoassay (CLIA). All hormonal measurements were performed in the hospital's central laboratory under standardized conditions using Elecsys Vitamin D Total Assay Kit (Roche Diagnostics, USA) and Access AMH Assay Kit (Coulter Beckman, USA).

Data were analyzed using SPSS software version 20.0. Continuous variables were expressed as mean ± standard deviation or median where appropriate, and categorical variables were expressed as percentages. The differences between groups were assessed using Student's t-test for normally distributed data and the Mann–Whitney U-test for non-normally distributed data. Chi-square or Fisher's exact test was used for categorical data. Logistic regression analysis was employed to identify independent predictors of clinical pregnancy. A p-value <0.05 was considered statistically significant.

The overall clinical pregnancy rate for IVF/ICSI clinical pregnancies was 37.8% (71/188). There was no statistically significant difference between the successful and unsuccessful pregnancy groups in terms of BMI, type of infertility, basal sex hormones bFSH, bLH, bE2, P, PRL, and T, number of Gn days, HCG day E2 (estrogen value), method of fertilization (IVF/ICSI), and oocyte-stage embryos (p > 0.05) (Table 1). In contrast, patients in the pregnancy group were significantly younger (mean age 29.89 ± 4.03 years) compared to the non-pregnancy group (mean age 32.42 ± 5.11 years; p = 0.026) and had shorter infertility durations (mean time 4.53 ± 2.15 vs. 5.32 ± 2.31 years; p = 0.011). These findings align with those of prior studies suggesting that advanced age and prolonged infertility negatively impact ART outcomes, likely due to declining ovarian reserve and reduced oocyte quality. Higher mean levels of AFC (17.25 ± 5.98 vs. 13.25 ± 6.62; p = 0.009) and AMH (3.68 ± 2.58 vs. 2.99 ± 1.41; p = 0.04) in the pregnancy group further support the importance of robust ovarian reserve in achieving successful ART outcomes. In addition, mean endometrial thickness on the day of HCG administration was greater in the pregnancy group (12.63 ± 2.09 mm vs. 10.52 ± 2.54 mm; p = 0.035) and mean vitamin D levels were higher (29.77 ± 13.3 vs. 24.97 ± 11.16 µg/L; p = 0.03), suggesting enhanced endometrial receptivity and potentially improved implantation potential.

To better understand the independent factors influencing ART pregnancy success, a stepwise backward logistic regression analysis was conducted. As presented in Table 2, seven independent predictors of clinical pregnancy were identified. These included age [β = 0.0921, odds ratio (OR) 1.08, 95% confidence interval (CI) 0.87–1.34; p = 0.016], AFC (β = 1.0652, OR 2.51, 95% CI 2.00–4.93; p = 0.033), vitamin D levels (β = 0.3758, OR 1.46, 95% CI 1.04–2.03; p = 0.027), AMH (β = 0.1697, OR 2.03, 95% CI 1.63–3.21; p = 0.037), Gn dose (β = 0.3296, OR 1.26, 95% CI 1.02–1.55; p = 0.030), endometrial thickness on HCG day (β = 1.6201, OR 1.97, 95% CI 1.28–2.04; p = 0.019), and the number of eggs retrieved (β = 1.5634, OR 1.44, 95% CI 0.48–1.62; p = 0.035).

Notably, each unit increased in vitamin D levels was associated with a 46% higher likelihood of clinical pregnancy, emphasizing its potential role in improving ART outcomes. Similarly, higher AFC and AMH levels strongly predicted pregnancy success, reflecting the critical role of ovarian reserve in achieving positive ART outcomes. In addition, greater endometrial thickness on the day of HCG and a higher number of eggs retrieved were positively associated with clinical pregnancy, suggesting the importance of endometrial receptivity and effective ovarian stimulation.

A full-variable logistic regression model was constructed to predict the probability of clinical pregnancy, with clinical pregnancy status as the dependent variable. The independent variables included factors that were significantly associated with pregnancy outcomes in previous analyses: age, AFC, vitamin D level, AMH, Gn dose, endometrial thickness on the day of HCG administration, and the number of eggs retrieved (Table 3). The predictive model for the probability of clinical pregnancy was expressed as: P = 1/(1 + y), where y = exp[−(3.538–0.0733  ×  age + 0.2962 × AFC + 0.4214 × vitamin D + 1.3624 × AMH + 0.2287 × Gn dosage + 0.2698 × endometrial thickness + 0.3241 × number of eggs retrieved)].This model integrates key clinical and biochemical factors to estimate the likelihood of clinical pregnancy in ART patients. It underscores the importance of combining age, ovarian reserve markers (AFC, AMH), vitamin D levels, and endometrial receptivity metrics to improve outcome prediction. The inclusion of vitamin D in the model highlights its potential role as a modifiable factor in enhancing ART success.

The predicted probability of clinical pregnancy was calculated using a regression model compared with actual clinical pregnancy outcomes to generate a ROC curve, determining the area under the ROC curve (AUC). The AUC of the model was 75.34% (standard error 0.045, 95% CI 0.712–0.965; p = 0.003), which indicated that the effectiveness of the logistic prediction model was good (Figure 1).

Among the 188 cases included in this study, 116 were classified into the vitamin D non-deficient group (vitamin D >50 nmol/L) and 72 were categorized into the vitamin D-deficient group. There were no significant differences between the two groups among baseline characteristics, including age, BMI, duration of infertility, type of infertility, ovarian reserve markers (AFC, AMH), basal sex hormone levels, COH regimens, Gn dosage, or fertilization methods (p > 0.05) (Table 4).

In contrast, several ART outcome indicators showed statistically significant differences between the two groups. Endometrial thickness on the day of HCG administration was significantly higher in the vitamin D non-deficient group compared to the deficient group (11.97 ± 1.65 mm vs. 10.07 ± 1.01 mm; p = 0.036), indicating better endometrial receptivity in patients with sufficient vitamin D levels. In addition, the number of embryos obtained was greater in the non-deficient group (5.36 ± 0.77 vs. 4.88 ± 0.97; p = 0.029), and the clinical pregnancy rate was significantly higher in the vitamin D non-deficient group (42.2% vs. 30.6%; p = 0.015). These findings suggest that sufficient vitamin D levels may improve ART outcomes by enhancing both endometrial receptivity and embryo quality.

As shown in Table 5, three factors associated with vitamin D levels were significantly related to pregnancy outcomes: the number of embryos obtained (β = 0.5532, OR 0.53, 95% CI 0.31–0.95; p = 0.024), endometrial thickness on the day of HCG administration (β = 0.2354, OR 1.97, 95% CI 0.44–1.68; p = 0.022), and clinical pregnancy (β = 0.3695, OR 0.55, 95% CI 0.33–1.24; p = 0.019). Patients with higher vitamin D levels had a significantly greater number of embryos obtained, as indicated by the positive association in the regression analysis. Similarly, endometrial thickness on the day of HCG was significantly greater in the vitamin D non-deficient group, suggesting a correlation between sufficient vitamin D levels and improved endometrial receptivity. Furthermore, vitamin D sufficiency was significantly associated with clinical pregnancy, highlighting its role in promoting implantation success.

Despite these promising findings, our study has several limitations. First, the single-center, retrospective design may restrict the generalizability of our results to other populations and clinical settings. Larger, multicenter prospective studies are necessary to validate our predictive model and ensure its applicability across diverse populations (11). Second, seasonal variation in serum 25(OH)D levels was not accounted for in our study. Vitamin D levels are known to fluctuate with seasonal sun exposure, which could have influenced the baseline vitamin D status of participants. For example, patients evaluated during winter months may have had lower serum 25(OH)D levels compared to those assessed in summer, potentially introducing variability in our results (20). Future studies should incorporate seasonal data to control for this factor and determine whether seasonal fluctuations impact ART outcomes. Third, although we analyzed vitamin D levels, the lack of comprehensive nutritional data, such as dietary intake and overall nutritional status, limits our ability to assess the broader impact of nutrition on ART outcomes. This limitation is particularly relevant given that overall nutrition, alongside vitamin D levels, plays a crucial role in reproductive health (8–10). In addition, interventional studies evaluating the effects of vitamin D supplementation on ART success are crucial to establish causality, determine optimal dosing protocols, and identify subpopulations of patients who may benefit most from supplementation. Furthermore, our findings suggest that vitamin D impacts both ovarian reserve and endometrial receptivity, emphasizing the need to investigate its genetic and molecular pathways in reproductive tissues. Finally, exploring interactions between vitamin D and other nutritional or hormonal factors may help refine holistic approaches to reproductive health, ultimately improving ART outcomes.

In conclusion, our study enhances the understanding of vitamin D's role in ART success and introduces a predictive model that incorporates both traditional and novel predictors of clinical pregnancy. It encourages a shift toward more holistic treatment paradigms in reproductive medicine, where nutritional and hormonal factors are considered in concert to optimize patient outcomes.