Lack of Association between Vitamin D Insufficiency and Cardiovascular or Fracture Risk: A UK Biobank Study

Article information

Endocrinol Metab. 2026;41(2):308-318

Publication date (electronic) : 2025 October 15

doi : https://doi.org/10.3803/EnM.2025.2482

Yongin Cho^1,*, Jong Hyun Jhee^2,*, Jong Ho Jhee³, Hye-Sun Park^,4

¹Department of Endocrinology and Metabolism, Inha University College of Medicine, Incheon, Korea

²Division of Nephrology, Department of Internal Medicine, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Korea

³Inria, Inserm, Paris Cité University (Université Paris Cité), UMR 1346, Paris, France

⁴Division of Endocrinology, Department of Internal Medicine, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Korea

Corresponding author: Hye-Sun Park. Department of Internal Medicine, Gangnam Severance Hospital, Yonsei University College of Medicine, 211 Eonju-ro, Gangnam-gu, Seoul 06273, Korea Tel: +82-2-2019-3313, Fax: +82-2-3463-3882, E-mail: hspark01@yuhs.ac

*These authors contributed equally to this work.

Received 2025 June 2; Revised 2025 July 14; Accepted 2025 August 1.

Abstract

GRAPHICAL ABSTRACT

INTRODUCTION

Vitamin D plays a crucial role in bone health by regulating calcium and phosphate metabolisms [1]. Beyond its skeletal benefits, vitamin D is also implicated in cardiovascular (CV) health through its effects on vascular inflammation, endothelial function, and the renin-angiotensin-aldosterone system [2]. Observational studies have consistently linked vitamin D deficiency to an increased risk of osteoporosis, fractures, bone loss, CV events, and mortality [2-4].

While a 25-hydroxy-vitamin D (25(OH)D) level of ≥50 nmol/L is widely accepted as the treatment goal for skeletal outcomes, some studies have suggested that higher levels (≥75 nmol/L) optimize the risk reduction of various endpoints [5,6]. The Endocrine Society once defined 50–75 nmol/L as ‘insufficient’ and ≥75 nmol/L as ‘sufficient’ [7], but retracted the insufficiency range in 2024 owing to insufficient evidence in healthy populations [8]. Additionally, the National Institutes of Health and Institute of Medicine consider ≥50 nmol/L adequate, citing limited bone health benefits beyond this level [9,10].

The lack of consensus on vitamin D targets stems from the inconsistent findings in existing studies on the association between vitamin D levels and health outcomes. For example, vitamin D deficiency was associated with an increased risk of hip fracture in women but only in the vitamin D range of <47.5 nmol/L [11]. In addition, vitamin D supplementation has yielded inconsistent results across various studies. A meta-analysis of randomized controlled trials suggested that daily supplementation with 700 to 800 IU of vitamin D reduced the risk of hip and non-vertebral fractures in older adults [12]. However, these benefits appear to be less evident in younger populations and healthier older adults. Similarly, the evidence on the benefits of vitamin D supplementation on CV health is inconclusive. While some meta-analyses suggest that vitamin D supplementation has protective effects against heart failure in older populations [13], large-scale trials such as the VITamin D and OmegA-3 TriaL (VITAL) have found no significant reduction in the risk of myocardial infarction (MI) or stroke [14]. Furthermore, Mendelian randomization (MR) analyses have shown no causal relationship between vitamin D concentration and CV events or mortality [15,16]. Given these inconsistencies, the clinical relevance of the historical insufficient range (50 to 75 nmol/L) remains uncertain. Therefore, this study aimed to investigate the association between increased fracture risk and CV outcomes using data from a large prospective cohort.

METHODS

RESULTS

The baseline characteristics of the study population by sex and vitamin D status are presented in Table 1 and Supplemental Table S2. In both women and men, lower vitamin D levels were associated with higher BMI, higher eGFR, and a higher proportion of current smokers. Participants in the deficient group also showed a lower prevalence of current alcohol consumption and a higher prevalence of hypertension, and diabetes (all P<0.001). Additional baseline characteristics used as covariates—including season of blood collection, skin color, sun exposure duration, geographic latitude, and calcium supplementation—are summarized in Supplemental Table S1. These factors showed significant differences across vitamin D status groups in both sexes (all P<0.001), reflecting known environmental and behavioral influences on vitamin D levels.

Table 1.

Baseline Characteristics of Study Participants

CV outcomes

During a median follow-up of 13.6 years, 23,527 cases of 3P-MACE were reported (Table 2). The incidence rates of 3P-MACE were 4.40 cases per 1,000 person-years in the vitamin D-sufficient group (≥75 nmol/L), 4.45 cases per 1,000 person-years in the insufficient group (≥50 to <75 nmol/L), and 5.08 cases per 1,000 person-years in the deficient group (<50 nmol/L), respectively. In the Cox proportional hazards model, vitamin D deficiency was associated with a 17% increased risk of 3P-MACE in the fully adjusted model (adjusted HR [aHR], 1.17; 95% CI, 1.11 to 1.24). A similar pattern was observed for specific 3P-MACE outcomes, with vitamin D deficiency linked to significantly high risks of MI (aHR, 1.13; 95% CI, 1.05 to 1.20), stroke (aHR, 1.21; 95% CI, 1.09 to 1.33), and CV mortality (aHR, 1.50; 95% CI, 1.32 to 1.70) (Supplemental Table S3). Within the deficient group, a 1-SD decrease of vitamin D was associated with 22% increased risk of 3P-MACE (aHR, 1.22; 95% CI, 1.17 to 1.28). In contrast, the vitamin D-insufficient group showed no significant association with 3P-MACE, MI, or stroke. However, vitamin D insufficiency was linked to a modestly increased risk of CV mortality (aHR, 1.15; 95% CI, 1.01 to 1.32), although the effect size was smaller than that observed in the deficient group (aHR, 1.50; 95% CI, 1.32 to 1.70) (Supplemental Table S3).

Table 2.

Association between Vitamin D Status and Risk of 3P-MACE

When stratified by sex, similar results were observed. Vitamin D deficiency was associated with an increased risk of 3P-MACE in both women (aHR, 1.12; 95% CI, 1.01 to 1.24) and men (aHR, 1.19; 95% CI, 1.12 to 1.27). Furthermore, within the deficiency group, each 1-SD decrease in vitamin D level was associated with a 19% higher risk of 3P-MACE in women (aHR, 1.19; 95% CI, 1.09 to 1.29) and a 23% higher risk in men (aHR, 1.23; 95% CI, 1.17 to 1.30). In contrast, no significant association was found in the insufficiency group for both women and men.

Fracture outcomes

During a median follow-up of 13.6 years, a total of 10,544 cases of MOFs were reported (Table 3). The incidence rates of MOFs were 2.22, 2.09, and 2.13 cases per 1,000 person-years in the sufficient, insufficient, and deficient groups, respectively. In the Cox proportional hazards model, individuals in the vitamin D-deficient group had a 9% higher risk of MOF compared to those in the sufficient group (aHR, 1.09; 95% CI, 1.01 to 1.18), whereas no significant risk increase was observed in the insufficient group (aHR, 0.98; 95% CI, 0.91 to 1.06) (Table 3). Furthermore, within the deficient group, each 1-SD decrease in vitamin D level was associated with an 18% increased risk of MOF (aHR, 1.18; 95% CI, 1.10 to 1.27).

Table 3.

Association between Vitamin D Status and Risk of MOF

We performed sex-stratified analyses. In men, similar patterns were observed. Those in the vitamin D-deficient group had a 30% higher risk of MOF compared to those in the sufficient group (aHR, 1.30; 95% CI, 1.12 to 1.50). Furthermore, within the deficient group, each 1-SD decrease in vitamin D level was associated with a 32% increased risk of MOF (aHR, 1.32; 95% CI, 1.16 to 1.48). Although men in the insufficiency group did not show a significantly increased risk of MOF compared to the sufficient group, a 1-SD decrease in vitamin D within the insufficiency group was associated with a 34% higher risk of MOF (aHR, 1.34; 95% CI, 1.07 to 1.67). In women, those in the vitamin D-deficient group did not show an increased risk of MOF compared to the sufficient group. However, within the deficient group, each 1-SD decrease in vitamin D level was associated with a 10% higher risk of MOF (aHR, 1.10; 95% CI, 1.01 to 1.20). In contrast, no significant association was observed between vitamin D levels and MOF within the insufficiency group.

Risk patterns across vitamin D categories

To further characterize the relationship between vitamin D status and clinical outcomes, we subdivided the deficiency group into moderate and severe deficiency, resulting in four categories: sufficient (≥75 nmol/L), insufficient (≥50 to <75 nmol/L), moderate deficiency (≥30 to <50 nmol/L), and severe deficiency (<30 nmol/L) (Fig. 1). Compared to the sufficient group, individuals in the severe deficiency group had significantly higher risks of 3P-MACE (aHR, 1.29; 95% CI, 1.21 to 1.37) and MOF (aHR, 1.20; 95% CI, 1.10 to 1.32). The moderate deficiency group showed a modest increase in the risk of 3P-MACE (aHR, 1.13; 95% CI, 1.07 to 1.20), but no significant association with MOF (aHR, 1.05; 95% CI, 0.97 to 1.14). No significant risk elevation was observed in the insufficient group for either outcome (aHR, 1.04; 95% CI, 0.99 to 1.10 for 3P-MACE; and aHR, 0.98; 95% CI, 0.91 to 1.06 for MOF). We further examined the association of continuous vitamin D levels with outcomes using spline curve analyses (Fig. 2). The curves showed a clear negative correlation between vitamin D levels and the risk of both 3P-MACE and MOF, but only below 50 nmol/L (Fig. 2). The steepest increase in risk was observed in the range of severe deficiency, while no significant association was noted within the range of vitamin D insufficiency.

Fig. 1.

Hazard ratios for (A) three-point major adverse cardiovascular events and (B) major osteoporotic fracture across vitamin D categories. Adjusted for age, sex, body mass index, education, smoking, alcohol, steroid use, activity status, history of cardiovascular disease, history of fracture, ethnicity, house income, estimated glomerular filtration rate, hypertension, diabetes mellitus, rheumatoid arthritis, and hyperthyroidism.

Fig. 2.

Spline curves for vitamin D levels and hazard ratios for (A) three-point major adverse cardiovascular event (3P-MACE) and (B) major osteoporotic fracture (MOF). The blue line represents the adjusted hazard ratio for 3P-MACE and MOF. The gray area indicates the 95% confidence interval. Curves were adjusted for age and sex. 25(OH)D, 25-hydroxy-vitamin D.

DISCUSSION

In this large population-based cohort study using UK Biobank data, we found that vitamin D deficiency (<50 nmol/L) was significantly associated with increased risks of both CV events (aHR, 1.17; 95% CI, 1.11 to 1.24) and MOF (aHR, 1.09; 95% CI, 1.01 to 1.18), compared to vitamin D sufficiency (≥75 nmol/L). Notably, individuals with severe deficiency (<30 nmol/L) had the highest risk estimates for both outcomes. In contrast, vitamin D insufficiency (50 to 75 nmol/L) was not associated with a statistically significant increase in risk. These findings suggest that the clinical relevance of the insufficiency range may be limited, supporting recent guideline changes that de-emphasize this category in routine screening and treatment strategies [8].

The association between vitamin D deficiency and various health outcomes has been extensively studied. In a study of 1,739 participants, individuals with 25(OH)D <37.5 nmol/L had a 62% higher risk for CV events compared to those with 25(OH) D ≥37.5 nmol/L [2]. The risk appeared graded, with individuals having 25(OH)D levels <25 nmol/L showing an 80% increased CV risk compared to those with levels ≥37.5 nmol/L. Supporting this, a systematic review reported that lower 25(OH)D concentrations were associated with incident CV disease in five of seven analyses across six cohorts [19]. Similarly, numerous studies have explored the relationship between vitamin D deficiency and fracture risk. van Schoor et al. [19] reported that individuals aged 65 to 75 years with low 25(OH)D levels (<30 nmol/L) had 3.1-fold increased fracture risk (HR, 3.1; 95% CI, 1.4 to 6.9). In a meta-analysis including 61,744 elderly population, the lowest categories of vitamin D had an odds ratio of 1.80 for hip fracture (odds ratio, 1.80; 95% CI, 1.56 to 2.07) [20]. These findings are consistent with our study, which also showed a strong association between vitamin D deficiency and the risks of 3P-MACE and MOF.

In our study, sex-stratified analyses revealed that vitamin D deficiency was significantly associated with MOF risk in men, but not in women with vitamin D deficiency. This result may be due to the greater influence of other fracture-related factors in women, such as bone mineral density, age at menopause, and history of medications affecting bone health. Furthermore, women are more likely to initiate vitamin D supplementation or begin anti-osteoporotic treatment during the follow-up period, which could have attenuated the observed association between baseline vitamin D status and fracture risk.

In our sensitivity analyses, vitamin D deficiency remained significantly associated with both 3P-MACE and MOF among individuals without a prior history of CV disease or fracture. While similar trends were observed in those with a prior history, the associations did not reach statistical significance. However, the P values for interaction were >0.999, suggesting no meaningful effect modification by prior history. Thus, the lack of statistical significance in the history-positive group likely reflects reduced power due to smaller sample size and fewer outcome events, rather than a true biological difference.

The relevance of vitamin D insufficiency and clinical outcome is less clearly defined. In a study of 1,045 Swedish women aged 75 years, those with intermediate vitamin D levels (50 to 75 nmol/L) had a higher 10-year fracture incidence compared to those with sufficient levels (>75 nmol/L) (21.3% vs. 12.3%, P=0.035) [21]. In contrast, another study reported that the increased fracture risk associated with low vitamin D levels was observed only below 40 nmol/L, with no increase in risk observed above this threshold [22]. This inconsistency may stem from several factors, including challenges in accurately measuring serum 25(OH)D, seasonal and geographic variation in vitamin D levels, and individual differences in sun exposure, diet, skin pigmentation, and body composition [23]. Additionally, assay variability across laboratories further complicates interpretation of borderline values. Given these confounders, establishing a universal threshold for defining insufficiency remains difficult, and the clinical implications of moderately low vitamin D levels may differ depending on population context and baseline risk.

In our study, we defined vitamin D insufficiency as 25(OH)D levels between 50 and 75 nmol/L, following the Endocrine Society’s guidelines [7]. We accounted for a wide range of factors that may influence vitamin D status in our analyses including seasons of blood draw, skin color, and geographic latitude. Unlike the vitamin D deficiency group, which showed increased risks of 3P-MACE and fractures, the vitamin D insufficiency group showed no significant association with either outcome. These findings were robust across various analytic approaches, suggesting that moderately low vitamin D levels may not confer increased risk in this population. Our results align with prior analyses from the UK Biobank, which reported associations between vitamin D deficiency and mortality, with most of the excess risk concentrated at levels below 50 and 60 nmol/L [24-26]. In particular, MR studies have provided evidence supporting a causal relationship between vitamin D deficiency and increased mortality risk, particularly at levels below 50 nmol/L. An L-shaped association has been observed in several studies, indicating that the risk rises steeply in the deficient range but plateaus beyond that threshold [27]. Other studies, however, did not identify a causal relationship across the full range of vitamin D levels [28], reflecting continued debate within the field.

Given that vitamin D deficiency is associated with unfavorable health outcomes, numerous studies demonstrated the vitamin D supplementation and outcomes, showing inconsistent results. A previous meta-analysis reported that high-dose vitamin D (≥800 IU/day) offered some protection against fractures in individuals aged 65 years or older [29], a population that is generally susceptible to vitamin D deficiency. In contrast, recent trial questioned the broad benefits of vitamin D supplementation, showing no significant reduction in fracture incidence with vitamin D supplementation in generally healthy middle-aged and older adults not selected for vitamin D deficiency [30]. Based on these findings, the 2024 Endocrine Society recommendation discourage routine screening or supplementation in the general population [8].

In our study, the effect of a 1-SD decrease in 25(OH)D levels within each vitamin D category may offer a rough estimate of the potential clinical benefit of supplementation. In the vitamin D deficiency group, a 1-SD decrease was associated with a 22% increased risk of 3P-MACE and an 18% increased risk of MOF. In contrast, within the insufficiency group, a 1-SD decrease was associated with only a modest 9% increase in 3P-MACE risk and had no significant impact on MOF. These results suggest that changes in vitamin D levels have greater clinical relevance within the deficiency range than within the insufficiency range. From a clinical standpoint, this may imply that individuals with vitamin D deficiency are more likely to benefit from supplementation, whereas the benefit in those with moderate insufficiency may be limited. Although large randomized trials such as VITAL and D-Health have shown no overall benefit of vitamin D supplementation in the general population [31,32], our findings suggest that individuals with confirmed deficiency remain at elevated risk and may warrant targeted intervention and further investigation in future trials.

Strengths of the study include the use of a large, well-characterized population-based cohort with over 380,000 participants, enabling sufficient power to detect clinically meaningful differences across vitamin D categories. We placed specific focus on the vitamin D insufficiency range, a category that has received limited attention and been inconsistently addressed in prior studies. Additionally, our integration of both CV and fracture outcomes provides a comprehensive perspective on the health implications of vitamin D status, and the use of spline analyses allowed for a nuanced assessment of potential threshold effects.

Despite its strengths, our study has several limitations. First, although individuals taking vitamin D supplements at baseline were excluded, we lacked data on supplementation or nutritional intake during the follow-up period. Additionally, serum 25(OH)D was measured only once at baseline, making it difficult to determine whether participants maintained the same vitamin D status throughout the study. However, previous research has demonstrated moderate-to-high within-individual stability of serum 25(OH)D levels over time; for example, Jorde et al. [33] reported a correlation coefficient of 0.80 between baseline and 12-month measurements in a 1-year intervention study, supporting the relevance of a single measurement in prospective analyses. Moreover, we minimized residual confounding by adjusting for multiple covariates including geographic latitude, sun exposure, skin color, and season at blood draw. Second, as an observational study, our findings cannot establish causality and remain subject to residual confounding and potential reverse causation. For example, individuals with subclinical disease at baseline may have had both lower vitamin D levels and a higher risk of subsequent events. To address this, we performed a sensitivity analysis excluding events that occurred within the first year of follow-up; the associations remained robust, lending support to the validity of our findings. Future studies incorporating repeated measurements of vitamin D and randomized trials targeting individuals with confirmed deficiency are warranted to better establish causality and assess the therapeutic benefit of supplementation. Third, baseline bone mineral density data were not available, which is an important limitation when evaluating the association between vitamin D status and fracture risk. Fourth, the generalizability of our findings may be limited by the demographic characteristics of the UK Biobank cohort, which includes mostly White individuals aged 40 to 69 years and tends to be healthier than the general population. In addition, the potential for survivor bias cannot be ruled out, particularly given the exclusion of individuals with conditions affecting calcium metabolism.

In conclusion, elevated risks of fractures and CV events were observed only among individuals with vitamin D deficiency (<50 nmol/L), and the inverse association with vitamin D levels appeared to be limited to this range. In contrast, no significant associations were observed in the insufficiency range (50 to 75 nmol/L), suggesting limited clinical relevance of this category. These findings support recent guideline revisions that de-emphasize the insufficiency range, while highlighting the need for further research to determine whether individuals with confirmed deficiency may benefit from targeted interventions.

Supplementary Material

Article information Continued