Bacterial vaginosis (BV) is a common vaginal infection caused by an imbalance in the vaginal microbiota and is characterized by a decrease in the number of lactobacilli together with an overproliferation of anaerobic bacteria (e.g., Gardnerella vaginalis, Prevotella, and Atopobium vaginae) (1). This microecological imbalance causes the vaginal pH to rise, disrupting the normal acidic environment and leading to a range of symptoms (2). Typical symptoms in about 50% of patients are a thin, grayish-white, or yellowish discharge with a fishy odor. This is especially common after sexual intercourse (3, 4). BV can cause many complications, including an increased risk of sexually transmitted infections (5–8), such as human immunodeficiency virus (HIV), herpes simplex virus (HSV), Neisseria gonorrhoeae, and Chlamydia trachomatis, and cause genitourinary infections in women and adverse pregnancy outcomes (9–12). In addition, BV can have an impact on the mental health and social quality of women’s lives (13, 14).

The etiology of BV is not fully understood, and the recurrence rate is high. One study showed a recurrence rate of 58 percent within 12 months of standard treatment with metronidazole (15). This high recurrence rate may be due to a number of factors, including bacterial biofilm formation, antibiotic resistance, incomplete restoration of the vaginal micro-ecosystem, and unsafe sexual practices (16–19). Therefore, existing treatments have limited effectiveness in preventing recurrence of BV, and there is an urgent need to explore new treatment strategies.

Vitamin D is a fat-soluble nutrient that is synthesized by the skin, mainly through exposure to sunlight, but also through diet and supplementation, and has the function of promoting calcium and phosphorus absorption, supporting cardiovascular health, regulating cell growth and differentiation, and immunity (20, 21). The role of vitamin D in female reproductive health has recently attracted increasing attention. Studies have shown that serum vitamin D may influence the development and progression of BV in several ways, such as exerting direct antibacterial effects, preventing pathogen biofilm formation, regulating immune and inflammatory responses, maintaining the acidic vaginal environment, and reprogramming epigenetics and metabolism (22–25). Although several studies have focused on the relationship between serum vitamin D and the risk of BV, the results in the existing literature are inconsistent. Turner et al. conducted a randomized controlled trial that showed that high-dose vitamin D supplementation in addition to standard metronidazole therapy did not reduce BV recurrence (26). Subsequently, in their study of Zimbabwean women of childbearing age, they showed that serum vitamin D insufficiency was not associated with increased prevalence or incidence of BV (27). However, some studies support the idea that people with vitamin D deficiency or insufficiency are more likely to develop BV (28, 29), and that vitamin D supplementation may be beneficial in treating BV (30). Therefore, further studies are needed to test the association between serum vitamin D and BV.

Our study showed a non-linear negative association between serum vitamin D levels and the risk of BV, but this association was no longer present at serum vitamin D concentrations ≥ 63.1 nmol/L. This study systematically analyzed the causes of conflicting conclusions in previous observational studies, laying an evidence-based medical foundation for precise vitamin D supplementation strategies for people at high risk of BV.

We found that serum vitamin D deficiency was associated with an increased risk of BV, which is consistent with previous hypotheses about the mechanism by which vitamin D regulates mucosal immunity and maintains vaginal microecological balance. Holick et al. (33) pointed out that vitamin D promotes the expression of antimicrobial peptides by activating the vitamin D receptor (VDR), improves epithelial barrier function, and inhibits pathogen colonization. In addition, vitamin D may protect the vagina from bacterial infection by increasing the stability, proliferation and differentiation of stratified squamous epithelial cells through activation of the VDR, p-RhoA and p-Ezrin pathways, as suggested in other systems (34). Vitamin D can also induce the synthesis and secretion of insulin and increase the phosphorylation and inactivation of glycogen synthase kinase (inhibitor of glycogen synthesis) in adipose tissue, indirectly promoting glycogen production. This may alter vaginal glucose homeostasis, increasing the relative abundance of lactobacilli and lowering local pH (35–37).

Smooth curve fitting revealed a non-linear negative association between serum vitamin D and the risk of BV and identified a threshold effect of 63.1 nmol/L (log likelihood ratio <0.001). The inflection point at 63.1 nmol/L may reflect a biological saturation effect, with the inverse association between vitamin D and BV risk plateauing beyond this level. Its clinical relevance requires validation in prospective interventional trials. This threshold is significantly lower than the 75 nmol/L recommended by the Endocrine Society for bone health, suggesting that vitamin D’s role in mucosal immune protection may be saturated earlier than its effect on bone health. Specifically, when serum vitamin D levels are below 63.1 nmol/L, the risk of BV increases with decreasing vitamin D levels, while above this threshold the protective effect stabilizes. This difference may be due to the local metabolic characteristics of the vagina: 1α-hydroxylase expressed by vaginal epithelial cells can convert vitamin D in serum to the active form 1,25(OH)₂D, and it can effectively trigger the VDR signaling pathway at lower serum concentrations (32, 38, 39). However, the skeletal system has a higher requirement for vitamin D, and levels usually need to be maintained above 75 nmol/L to inhibit parathyroid hormone (PTH) secretion and maintain calcium and phosphorus homeostasis (32, 40). This finding provides an important basis for stratified intervention: for women at high risk for BV, the target for serum vitamin D optimization can be set at ≥63.1 nmol/L rather than following the standard for bone health uniformly. Maintaining serum vitamin D levels ≥63.1 nmol/L was associated with both a lower risk of bacterial vaginosis and a reduced likelihood of harm related to excessive supplementation. However, this threshold is an exploratory result that is not sufficient to establish clinical guidelines.

Subgroup analyses of this study showed that the negative association between serum vitamin D and the risk of BV remained robust (P-interaction > 0.05) in most stratifications, such as age, education level, and marital status, suggesting the broad biological effects. However, this link did not reach statistical significance in the subgroups certain ethnic groups and “BMI < 18.5.” This could be because of a small sample size or factors that were unique to those groups. It is worth noting that, as a fat-soluble vitamin, vitamin D is closely associated with adipose tissue in terms of metabolism, storage, and distribution (41). It’s harder to store vitamin D if there is a big loss of fat tissue, like when a person has a low BMI or cachexia (42). This could cause the results to be skewed. One possible explanation for the finding that the link was weaker in the “BMI < 18.5” subgroup is this mechanism.

The present study has the following limitations. First, while the biological plausibility of vitamin D’s immunomodulatory effects on vaginal microbiota remains temporally stable, our reliance on 2001–2004 NHANES data introduces potential contextual biases. Temporal shifts in environmental determinants—particularly evolving dietary patterns, sunscreen utilization trends, and lifestyle modifications—may alter population-level vitamin D status profiles. Furthermore, the cross-sectional design inherently precludes causal inference regarding vitamin D’s protective mechanisms against bacterial vaginosis (BV). Subsequent research should prioritize contemporary cohorts with standardized BV phenotyping to validate these findings in modern populations. Second, although the NHANES database reduces seasonal bias in vitamin D measurements by using a seasonally adjusted sampling scheme (e.g., balanced inclusion of samples tested in different seasons) (43), individual-level variations in sun exposure and dietary intake may still introduce residual confounding in the results. Third, there may have been a higher risk of type II error because the sample sizes for subgroups like “other races” (n = 57) and “BMI < 18.5” (n = 34) were not big enough. Fourth, even after accounting for important demographic and lifestyle factors, some unmeasured influences (like how often people have sex, use contraception, maintain vaginal hygiene, and their hormonal levels) might still affect the results (44). Future studies need to include prospective cohorts or randomised controlled trials (RCTs) to verify causality, while more confounders and dynamic monitoring of vitamin D levels are recommended.

In summary, this study supports an association between vitamin D insufficiency and an increased risk of BV, and this association is consistent in most subgroups of the population. The fact that this association is no longer significant when serum vitamin D concentrations reach or exceed 63.1 nmol/L provides an important basis for the development of precise intervention strategies.