Primary relevance to vitamin D Vitamin D is a stated subject of this paper.
Vitamin D is a stated subject of this paper; the supporting source evidence is shown below.
Why this paper is included
Supporting source evidence:
Restricted cubic spline regression model depicting the association between serum vitamin D (VitD) levels and the odds of ruptured intracranial aneurysms.
Source: Associations of Vitamin D With IA Rupture in the Observational Study. View original paper
Relevance classification does not establish study quality, causation or editorial approval.
Study design Genetic instrumental-variable study.
Genetic instrumental-variable study
Uses genetic variation as an instrument to investigate relationships; this is not a randomised treatment trial.
Meta-analysis unclear · Blinding unclear / not reported · Research setting awaiting verification
How this paper was classified
Automatic source-based classification; awaiting editorial verification. This is not a quality score.
Serum Vitamin D Levels Were Not Associated With the Risk of Aneurysmal Subarachnoid Hemorrhage: A Large Cohort Study With Propensity Score Matching and Mendelian Randomization Analysis
Source: Article title.
What the study types mean
- Systematic review
- Uses explicit methods to identify, assess and summarise research on a defined question.
- Scoping review
- Maps the extent and characteristics of research on a topic.
- Narrative review
- Discusses existing research; systematic-review methods have not been established by this classification.
- Review / evidence synthesis
- Reviews existing research; the exact review methods need further assessment.
- Randomised controlled trial
- Assigns participants or groups to comparison interventions by chance.
- Non-randomised intervention study
- Examines an intervention without random allocation.
- Cohort study
- Examines exposures and outcomes over time within a defined group.
- Case-control study
- Compares people with an outcome against people without it, examining exposures.
- Cross-sectional study
- Examines a population at a particular point in time.
- Observational study
- Observes exposures and outcomes without assigning the exposure; the subtype is not established here.
- Case report
- Describes an individual case.
- Case series
- Describes a group of individual cases.
- Secondary analysis
- Analyses existing study data; this paper is not automatically another independent trial.
- Laboratory study
- Investigates biological processes in a laboratory setting.
- Qualitative study
- Examines experiences or meanings using qualitative methods.
- Genetic instrumental-variable study
- Uses genetic variation as an instrument to investigate relationships; this is not a randomised treatment trial.
- Study protocol
- Describes planned methods and is not evidence that the planned study has reported results.
- Multiple reported designs
- More than one design is explicitly reported; examine the supporting methods.
- Study design awaiting verification
- The available source does not establish a sufficiently clear study design.
Meta-analysis combines numerical results from studies. Blinding describes who was unaware of treatment allocation. Neither is a separate quality grade. Study design, vitamin D relevance and document format are independent filters.
Full studies
Serum Vitamin D Levels Were Not Associated With the Risk of Aneurysmal Subarachnoid Hemorrhage: A Large Cohort Study With Propensity Score Matching and Mendelian Randomization Analysis
Genetic instrumental-variable study · CNS Neuroscience & Therapeutics · 2025
Original publication title
Serum Vitamin D Levels Were Not Associated With the Risk of Aneurysmal Subarachnoid Hemorrhage: A Large Cohort Study With Propensity Score Matching and Mendelian Randomization Analysis
© 2025 The Author(s). CNS Neuroscience & Therapeutics published by John Wiley & Sons Ltd.
Original publication · Publisher record · CC BY 4.0
Source wording retained. Formatting, section navigation and display-title punctuation adapted for this research record.
At a glance
Study background
What background does the paper provide?
Intracranial aneurysms (IAs) are abnormal dilations of the cerebral arterial wall, posing a significant health risk due to their potential rupture, which can cause subarachnoid hemorrhage (SAH), a life‐threatening condition with high morbidity and mortality [1]. Incidental IAs are increasingly diagnosed owing to the widespread use of noninvasive neuroimaging. However, not all unruptured IAs rupture during a patient’s lifetime. Clinical factors such as hypertension, smoking, history of SAH, as well as aneurysm size and location, have been associated with IA growth and rupture [2, 3].
Tools such as the ELAPSS and PHASES scores, which incorporate these factors, have been developed to predict the risk of IA growth and rupture [4, 5]. Nevertheless, the literature lacks appropriate hematological indices for predicting aneurysm rupture.
While the pathological mechanisms underlying IA rupture remain incompletely understood, factors such as inflammation, oxidative stress, hemodynamic stress, and genetic predispositions are considered contributing factors [2, 3]. Abnormal blood flow and shear stress may compromise vascular integrity, while inflammation and oxidative stress could further damage the aneurysm wall, thereby increasing the rupture risk [2, 3, 6]. Vitamin D (VitD), a fat‐soluble steroid vitamin essential for calcium homeostasis and immune regulation, may confer vascular health benefits due to its anti‐inflammatory and antioxidative properties [7, 8].
Low levels of VitD are associated with elevated vascular inflammation, oxidative stress, endothelial dysfunction, and increased risk of cardiovascular and cerebrovascular events, including hypertension, atherosclerosis, myocardial infarction, and stroke [7, 8, 9]. VitD deficiency is also prevalent among IA patients [10, 11]. However, its relationship with IA rupture remains unclear. One study reported an inverse correlation between VitD levels and the risk of aneurysmal subarachnoid hemorrhage (aSAH) [12], whereas another found no association [11].
These discrepancies may arise from variations in sample size, study populations, methodologies, or unaccounted confounders. Additionally, VitD deficiency could result from aneurysm rupture, suggesting potential reverse causality. Thus, the specific role of VitD in aSAH remains controversial, and further research is required to determine whether VitD deficiency is associated with IA rupture.
Therefore, in this study, a large case: control study with propensity score matching (PSM) to balance confounding factors was conducted to investigate the potential association between VitD levels and aSAH. However, observational studies remain vulnerable to unmeasured confounders. To address this limitation, we also performed a two‐sample Mendelian randomization (MR) analysis to explore potential causal relationships. By integrating these methods, this study aims to provide high‐quality evidence regarding the relationship between VitD status and aSAH. The findings may improve our understanding of IA rupture mechanisms.
Source: Introduction Source: Introduction Source: IntroductionStudy methods
What methods does the paper describe?
Observational Study Design
This retrospective case: control study comprised consecutive patients with IA treated at Qilu Hospital of Shandong University between March 2018 and May 2021. The study was approved by the Ethics Committee of Qilu Hospital of Shandong University, and written informed consent was obtained from all participants prior to enrollment. All IAs were confirmed using computed tomography (CT), magnetic resonance imaging (MRI), or digital subtraction angiography (DSA). SAH was confirmed via CT scans. Ruptured IAs were defined as cases presenting with aneurysmal SAH, regardless of the number of aneurysms present, while unruptured IAs were defined as patients with at least one aneurysm and no history of SAH.
Exclusion criteria were as follows: (1) recent administration of VitD supplements or glucocorticoids within 1 month prior to the study; (2) concomitant vascular diseases, such as arteriovenous malformations, moyamoya disease, or arteriovenous fistulas; (3) recurrent, fusiform, traumatic, bacterial, or dissecting aneurysms; (4) aneurysmal SAH occurring more than 72 h before admission; (5) SAH of unknown etiology or non‐aneurysmal SAH; and (6) age < 18 years. Ultimately, 499 patients with either unruptured or ruptured IAs were included in the study. The study flowchart is presented in Figure 1A.
The flow chart of the present study.
The patients’ demographic characteristics, comorbidities (including diabetes mellitus, hypertension, and hyperlipidemia), and aneurysm characteristics (such as location and number) were recorded. For patients with multiple aneurysms, the study evaluated the location (anterior or posterior circulation) of the primary aneurysm treated during hospitalization. Fasting venous blood samples were collected within 24 h of admission to measure 25‐hydroxyvitamin D levels and conduct lipid profile tests, including triglycerides (TG), total cholesterol (TC), low‐density lipoprotein cholesterol (LDL‐C), and high‐density lipoprotein cholesterol (HDL‐C). Body mass index (BMI) was categorized as underweight (≤ 18.5 kg/m2), normal weight (18.5 to 24 kg/m2), overweight (24 to 28 kg/m2), or obese (> 28 kg/m2).
Hypertension was defined as a documented history of the condition (systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg), irrespective of antihypertensive treatment. Diabetes mellitus was defined as patients receiving antidiabetic medication, a fasting glucose level ≥ 7 mmol/L from previous health assessments, or a hemoglobin A1c level ≥ 6.5%. Hyperlipidemia was defined as lipid‐lowering medication use or fasting plasma TC ≥ 6 mmol/L, TG ≥ 2 mmol/L, or LDL ≥ 3.5 mmol/L. Current smoking was defined as consumption of ≥ 5 cigarettes per day, and alcohol consumption as intake of ≥ 150 g per week.
VitD levels were categorized as deficient (< 20 ng/mL), insufficient (20 to 29 ng/mL), or normal (≥ 30 ng/mL).
Statistical analyses were performed using SPSS 25.0 (IBM Corp.). Continuous variables were expressed as mean ± standard deviation or median [interquartile range], depending on their distribution. The Shapiro: Wilk test was used to assess normality, with a p‐value < 0.05 indicating a deviation from normal distribution. Group comparisons were conducted using Student’s t‐test for normally distributed variables and the Mann: Whitney U test for non‐normally distributed variables. Categorical variables were expressed as frequencies and percentages and were compared using the χ 2 test or Fisher’s exact test.
Factors with p < 0.1 in univariate analysis were included in binary logistic regression analysis using an enter process to identify independent predictors. PSM was used to balance confounding factors between patients with and without aneurysmal SAH. A one‐to‐one PSM analysis using the nearest neighbor method with a caliper value of 0.01 was conducted to adjust for imbalanced factors. A restricted cubic spline (RCS) regression model was employed to examine potential linear or nonlinear associations between serum VitD levels and the risk of aSAH.
Four knots were positioned at the 5th, 35th, 65th, and 95th percentiles of the VitD distribution, with the median serving as the reference point. The RCS model was fitted using the “rms” package (version 6.8 to 2), and spline plots were generated using the “plotRCS” package (version 0.1.5) in R software (version 4.4.1). Statistical significance was defined as a two‐tailed p‐value < 0.05.
MR Analysis
A two‐sample MR analysis was performed to investigate the causal relationship between VitD levels and aSAH. Figure 1B shows the analysis flowchart.
Genetic instruments for VitD levels were obtained from genome‐wide association studies (GWAS) datasets available through the IEU OpenGWAS project (https://gwas.mrcieu.ac.uk/), a public resource for investigating genetic associations with complex traits. Summary‐level GWAS data on IAs were derived from two large cohorts: the discovery cohort, conducted by the International Stroke Genetics Consortium (ISGC) with individuals of European ancestry, and the replication cohort from FinnGen. The discovery cohort included 5140 ruptured cases, 2070 unruptured cases, and 71,934 controls, while the replication cohort had 6236 ruptured cases, 3014 unruptured cases, and 412,105 controls.
Additional details regarding the data sources used in this study are provided in File S1.
From these GWAS datasets, single nucleotide polymorphisms (SNPs) significantly associated with VitD levels (p < 5 × 10−8) were selected as candidate instrumental variables (IVs). Linkage disequilibrium (LD) analysis was conducted to ensure SNP independence, applying a threshold of LD‐r 2 < 0.001 and a clumping distance threshold of > 10,000 kb. The F‐statistic [F = R 2(N‐2)/(1‐R 2)] was calculated to evaluate IV strength, with SNPs failing to meet the criterion of F ≥ 10 excluded to mitigate weak instrument bias, thereby maintaining the robustness of the analysis.
The inverse variance‐weighted (IVW) method was employed as the primary approach for causal estimation, providing reliable results in the absence of heterogeneity or horizontal pleiotropy. Supplementary methods: including MR‐Egger regression, weighted median, simple mode, and weighted mode: were used to ensure robustness under varying conditions. Heterogeneity was assessed using Cochran’s Q test, and if heterogeneity was detected (p < 0.05), adjusted results were reanalyzed [13]. Horizontal pleiotropy was evaluated using the MR‐Egger intercept method, with p > 0.05 indicating no pleiotropy and supporting MR reliability [14].
The leave‐one‐out (LOO) method was applied to systematically exclude each SNP to assess result stability. All analyses were conducted using R software (version 4.4.1) with the “TwoSampleMR” package [15], with statistical significance defined as p < 0.05.
Associations of Vitamin D With IA Rupture in the Observational Study
Results: Associations of Vitamin D With IA Rupture in the Observational Study
The baseline characteristics are presented in Table 1. A total of 499 patients with 619 IAs were included in this study, comprising 337 (67.5%) females and 162 (32.5%) males. The median age was 58.0 years [52.0, 64.0], and the median BMI was 24.8 kg/m2 [22.3, 27.3]. Among these patients, 184 (36.9%) had ruptured IAs, and 105 (21.0%) had multiple IAs.
Baseline information of enrolled patients before or after propensity score matching (PSM).
| Characteristic | Before PSM | After PSM | ||||||
|---|---|---|---|---|---|---|---|---|
| Overall (n = 499) | Unruptured (n = 315) | Ruptured (n = 184) | p | Overall (n = 224) | Unruptured (n = 112) | Ruptured (n = 112) | p | |
| Sex (Female) | 337 (67.5) | 215 (43.1) | 122 (24.4) | 0.654 a | 157 (70.1) | 80 (35.7) | 77 (34.4) | 0.662 a |
| Age (year) | 0.026 a | 0.774 a | ||||||
| ≤ 50 | 110 (22.0) | 58 (11.6) | 52 (10.4) | 43 (19.2) | 24 (10.7) | 19 (8.5) | ||
| 51 to 60 | 178 (35.7) | 110 (22.0) | 68 (13.6) | 85 (37.9) | 41 (18.3) | 44 (19.6) | ||
| 61 to 70 | 168 (33.7) | 118 (23.6) | 50 (10.0) | 78 (34.8) | 37 (16.5) | 41 (18.3) | ||
| > 70 | 43 (8.6) | 29 (5.8) | 14 (2.8) | 18 (8.0) | 9 (7.3) | 10 (8.1) | ||
| BMI | 0.699 a | 0.625 b | ||||||
| Underweight | 15 (3.0) | 9 (1.8) | 6 (1.2) | 7 (3.1) | 3 (1.3) | 4 (1.8) | ||
| Normal | 188 (37.9) | 114 (22.8) | 75 (15.0) | 89 (39.7) | 48 (21.4) | 41 (18.3) | ||
| Overweight | 204 (40.7) | 134 (26.9) | 69 (13.8) | 94 (42.0) | 47 (21.0) | 47 (21.0) | ||
| Obese | 92 (18.4) | 58 (11.6) | 34 (6.8) | 34 (15.2) | 14 (6.3) | 20 (8.9) | ||
| Smoking | 85 (17.0) | 50 (10.0) | 35 (7.0) | 0.367 a | 36 (16.1) | 19 (8.5) | 17 (7.6) | 0.716 a |
| Drinking | 53 (10.6) | 29 (5.8) | 24 (4.8) | 0.180 a | 25 (11.2) | 11 (4.9) | 14 (6.3) | 0.524 a |
| Diabetes | 70 (14.0) | 53 (10.6) | 17 (3.4) | 0.019 a | 20 (8.9) | 9 (4.0) | 11 (4.9) | 0.639 a |
| Hypertension | 291 (58.3) | 184 (36.9) | 107 (21.4) | 0.955 a | 132 (58.9) | 63 (28.1) | 69 (30.8) | 0.415 a |
| Hyperlipidemia | 142 (28.5) | 91 (18.2) | 51 (10.2) | 0.780 a | 56 (25.0) | 28 (12.5) | 28 (12.5) | 1.000 a |
| Blood test dates | 0.147 a | 0.885 a | ||||||
| Mar: May | 130 (26.1) | 81 (16.2) | 49 (9.8) | 60 (26.8) | 28 (12.5) | 32 (14.3) | ||
| Jun: Aug | 100 (20.0) | 57 (11.4) | 43 (8.6) | 45 (20.1) | 23 (10.3) | 22 (9.8) | ||
| Sep: Nov | 156 (31.3) | 96 (19.2) | 60 (12.0) | 73 (32.6) | 36 (16.1) | 37 (16.5) | ||
| Dec: Feb | 113 (22.6) | 81 (16.2) | 32 (6.4) | 46 (20.5) | 25 (11.2) | 21 (9.4) | ||
| Anterior circulating aneurysm | 477 (95.6) | 301 (60.3) | 176 (35.3) | 0.960 a | 216 (96.4) | 110 (49.1) | 106 (47.3) | 0.280 a |
| Multiple aneurysms | 105 (21.0) | 56 (11.2) | 49 (9.8) | 0.019 a | 45 (20.1) | 21 (9.4) | 24 (10.7) | 0.617 a |
| Vitamin D (ng/mL) | 17.0 [12.1 to 22.4] | 17.6 [12.6 to 22.6] | 16.3 [11.6 to 22.3] | 0.418 c | 16.0 [11.5 to 21.3] | 15.8 [11.7 to 21.4] | 16.2 [10.8 to 21.2] | 0.904 c |
| Vitamin D levels | 0.339 a | 0.947 a | ||||||
| Normal | 46 (9.2) | 26 (5.2) | 20 (4.0) | 22 (9.8) | 11 (4.9) | 11 (4.9) | ||
| Insufficient | 133 (26.7) | 90 (18.0) | 43 (8.6) | 48 (21.4) | 23 (10.3) | 25 (11.2) | ||
| Deficient | 320 (64.1) | 199 (39.9) | 121 (24.2) | 154 (68.8) | 78 (34.8) | 76 (33.9) | ||
Note: Variables are presented as number of patients (%) or median (interquartile range), as appropriate.
p‐values calculated using Pearson’s chi‐square test.
p‐values calculated using Fisher’s exact test.
p‐values calculated using Mann: Whitney U test.
In this study, diabetes was inversely associated with the risk of aSAH (p = 0.034; odds ratio [OR], 0.53; 95% confidence interval [CI], 0.290 to 0.954), whereas multiple aneurysms increased the risk of aSAH (p = 0.006; OR, 1.88; 95% CI, 1.197 to 2.940) in multivariate analysis (Table S1). The prevalence of diabetes was 16.8% (53/315) in patients with unruptured IAs, compared with 9.2% (17/184) in patients with ruptured IAs. Conversely, multiple aneurysms were more prevalent in ruptured IA patients (26.6%, 49/184) than in unruptured IA patients (17.8%, 56/315).
Age, sex, BMI, smoking, alcohol use, hypertension, hyperlipidemia, and aneurysm location (anterior circulation) showed no significant association with aSAH in multivariate analysis (p > 0.05).
Blood test dates, as presented in Table 1, did not differ significantly between patients with unruptured IA and those with ruptured IA (p = 0.147). The median VitD level was 17.0 ng/mL [12.1 to 22.4]. VitD levels were normal in 9.2% of patients, insufficient in 26.7%, and deficient in 64.1% of patients. No significant difference in VitD levels was observed between patients with unruptured IA and those with ruptured IA (p = 0.418). Furthermore, VitD deficiency showed no association with aSAH in this study (p = 0.339).
To control for confounders, PSM was performed in a 1:1 ratio based on sex, age, BMI, diabetes, hypertension, hyperlipidemia, smoking, alcohol consumption, blood test date, and presence of multiple aneurysms. A total of 112 patients with ruptured IA were successfully matched to 112 patients with unruptured IA. Following PSM, no significant differences were observed in the distribution of the aforementioned variables. VitD levels and deficiency remained unassociated with aSAH (p > 0.05).
Additionally, RCS regression analysis showed no significant nonlinear association between serum VitD levels and the risk of IA rupture (p = 0.739), nor a significant overall association (p = 0.865), further supporting the findings before and after PSM (Figure 2).
Restricted cubic spline regression model depicting the association between serum vitamin D (VitD) levels and the odds of ruptured intracranial aneurysms.
The Causal Associations Between Vitamin D Status and IA Rupture in MR Analysis
Results: The Causal Associations Between Vitamin D Status and IA Rupture in MR Analysis
The IVW‐MR analysis revealed no causal relationship between VitD levels and the risks of aSAH (p = 0.966; OR, 1.00; 95% CI, 0.82 to 1.23) or unruptured IA (p = 0.74; OR, 0.95; 95% CI, 0.68 to 1.31) (Figure 3). These findings remained consistent across multiple MR methods, including weighted median, MR‐Egger, weighted mode, and simple mode approaches, and were replicated in an independent cohort (Table 2). Scatter plots and forest plots further confirmed these findings (Figures S1 and S2).
All SNPs used as instrumental variables (IVs) exhibited F‐statistics exceeding 20, indicating minimal weak instrument bias (Files S2: , S5). Sensitivity and heterogeneity analyses revealed no evidence of pleiotropy or heterogeneity (Table 3), supporting the robustness and reliability of the MR findings. Leave‐one‐out analysis further validated these results (Figure S3).
Forest plots for the causal associations of vitamin D (VitD) levels with ruptured intracranial aneurysms or unruptured intracranial aneurysms (UIAs) in discovery cohort.
The MR results of the replication cohort.
| Exposure | MR method | SNP | p | OR (95% CI) | |
|---|---|---|---|---|---|
| RIA | Vitamin D | MR Egger | 93 | 0.957 | 0.993 (0.777 to 1.270) |
| Weighted median | 93 | 0.417 | 0.891 (0.675 to 1.177) | ||
| Inverse variance weighted | 93 | 0.389 | 1.076 (0.911 to 1.270) | ||
| Simple mode | 93 | 0.745 | 0.907 (0.504 to 1.633) | ||
| Weighted mode | 93 | 0.948 | 1.007 (0.821 to 1.234) | ||
| UIA | Vitamin D | MR Egger | 93 | 0.410 | 0.874 (0.636 to 1.202) |
| Weighted median | 93 | 0.231 | 0.803 (0.561 to 1.150) | ||
| Inverse variance weighted | 93 | 0.371 | 1.103 (0.889 to 1.369) | ||
| Simple mode | 93 | 0.005 | 2.910 (1.413 to 5.990) | ||
| Weighted mode | 93 | 0.596 | 0.928 (0.706 to 1.221) |
Abbreviations: RIA, Ruptured intracranial aneurysms; UIA, Unruptured intracranial aneurysms.
Heterogeneity and pleiotropy tests of MR in discovery and replication cohort.
| Cohort | Test | Outcome | Exposure | Method | Effect size | p |
|---|---|---|---|---|---|---|
| Discovery | Heterogeneity | RIA | Vitamin D | Cochran’s Q test | 66.0 (QMR Egger) | 0.169 |
| 66.1 (QIVW) | 0.192 | |||||
| UIA | Vitamin D | Cochran’s Q test | 69.8 (QMR Egger) | 0.101 | ||
| 70.2 (QIVW) | 0.112 | |||||
| Pleiotropy | RIA | Vitamin D | MR‐Egger regression | 0.001 (Egger regression intercept) | 0.825 | |
| UIA | Vitamin D | MR‐Egger regression | −0.005 (Egger regression intercept) | 0.580 | ||
| Replication | Heterogeneity | RIA | Vitamin D | Cochran’s Q test | 110.8 (Q MR Egger) | 0.078 |
| 111.7 (Q IVW) | 0.080 | |||||
| UIA | Vitamin D | Cochran’s Q test | 88.6 (Q MR Egger) | 0.551 | ||
| 92.4 (Q IVW) | 0.469 | |||||
| Pleiotropy | RIA | Vitamin D | MR‐Egger regression | 0.003 (Egger regression intercept) | 0.393 | |
| UIA | Vitamin D | MR‐Egger regression | 0.009 (Egger regression intercept) | 0.055 |
Abbreviations: RIA, ruptured intracranial aneurysms; UIA, unruptured intracranial aneurysms.
Discussion
The relationship between VitD status and the risk of aSAH has received increasing attention due to accumulating evidence demonstrating the high prevalence of VitD deficiency in the population and the potential protective role of VitD in cardiovascular and cerebrovascular health [7, 8, 9, 16]. However, our study, which employed a large case: control design with PSM and two‐sample MR analysis, did not identify any association or causal relationship between VitD status and the risk of IA rupture.
VitD has a potential protective role in maintaining arterial health [17]. VitD receptors are present in vascular smooth muscle and endothelial cells. VitD deficiency is common in patients with IAs, even in those with aSAH and high sun exposure [18]. Although VitD might protect against IAs and reduce the risk of aSAH, our findings did not support this hypothesis, as VitD levels were not inversely associated with aSAH. A previous cross‐sectional study found higher VitD levels in patients with unruptured IAs than in those with ruptured IAs, with VitD independently associated with rupture risk [12].
However, that study had a small sample size and lacked PSM analysis. Our use of PSM balanced baseline characteristics between ruptured and unruptured IA groups, minimizing confounders such as age, sex, and comorbidities. Despite this, we found no significant difference in VitD levels between patients with ruptured or unruptured IA, and VitD deficiency was not associated with aSAH. Our results were consistent with a previous study [10], which found no difference in VitD levels between ruptured and unruptured IA patients, despite higher rates of hypovitaminosis D in IA patients requiring treatment compared to patients without IA.
However, our observational study may still be affected by uncontrolled confounders, such as lifestyle, dietary habits, sunlight exposure, and diseases or medications influencing VitD metabolism, which could potentially impact serum VitD levels and our findings. Future studies incorporating detailed lifestyle and environmental data are warranted to better account for these influences.
To address this limitation, we performed MR analysis, which utilizes genetic variants as instrumental variables to infer causal relationships. This analysis revealed no causal association between genetically determined VitD levels and IA rupture, consistent with prior MR studies utilizing distinct GWAS cohorts that also found no causal effect of VitD on IA rupture. We further used a restricted cubic spline regression model to evaluate potential nonlinear associations between serum VitD levels and IA rupture risk. The analysis revealed no significant linear or nonlinear relationship, thereby reinforcing our findings.
These results suggest that previously reported associations may be attributable to residual confounding or unmeasured biases. Although lower VitD levels were frequently observed in aneurysm patients, a recent meta‐analysis of VitD intervention trials failed to demonstrate a significant reduction in cardiovascular events with VitD supplementation, suggesting that lower VitD levels may represent an epiphenomenon rather than a causative factor [19]. One recent study reported VitD deficiency could promote IA rupture in a murine model [20]. However, all SAH murine models have inherent limitations, and the specific animal model used in that study may not fully replicate human aneurysm rupture mechanisms.
Therefore, these findings cannot be directly extrapolated to humans.
The biological mechanisms through which VitD influences vascular health may not directly relate to the pathophysiology of IA rupture. Although VitD may contribute to general vascular maintenance, the specific processes leading to aneurysm formation and rupture: such as hemodynamic stress, vascular wall degeneration, and inflammatory infiltration: may be less affected by VitD levels. Additionally, the VitD threshold for vascular protection may be lower than deficiency levels, meaning even “low” VitD levels could suffice for vascular integrity. However, studies indicate that VitD has anti‐inflammatory effects [21], and cerebrovascular inflammation is a major pathological factor for aneurysm rupture.
Low VitD levels are also associated with risk factors for aSAH, such as hypertension, hyperlipidemia, and hyperhomocysteinemia [22]. Therefore, our findings should be interpreted with caution. An alternative explanation is that the relationship between VitD and IA rupture is more complex than a linear association. VitD may directly or indirectly influence measurable and unmeasurable factors that contribute to the pathological progression of aneurysms. Its vascular effects might also be more pronounced in specific subgroups, such as individuals with certain comorbidities or genetic predispositions, and these effects may have been attenuated in our overall study population.
Although our study did not identify a significant association between VitD levels and IA rupture, this finding does not rule out VitD’s potential role in other aspects of aneurysm biology, such as formation or progression. Future studies should investigate these areas through longitudinal study designs or by focusing on subgroups more susceptible to the effects of VitD. Further research should also explore interactions between VitD and other factors, such as genetic variation, environmental exposures, or comorbidities, which may modulate IA rupture risk.
For instance, VitD might exert protective effects only in individuals with specific genetic profiles or under particular environmental conditions. Elucidating these interactions could provide deeper insights into VitD’s role in vascular pathophysiology.
The strengths of our study include the utilization of PSM and MR analysis to mitigate confounding and establish causal inferences. PSM effectively balances control groups and reduces selection bias in observational studies, while MR reliably evaluates causality by employing genetic variants as instrumental variables for VitD levels, thereby minimizing confounding and reverse causation. However, the MR analysis design differs from our observational analysis. While the observational study compared VitD levels in patients with ruptured and unruptured IA, the MR analysis assessed genetically predicted VitD levels in relation to unruptured IA and aSAH risk versus general population controls.
Nevertheless, these complementary methods provide both real‐world and genetically informed insights into the potential role of VitD in the pathophysiology of IA.
However, our study has several limitations. First, while PSM balanced observed baseline characteristics, unmeasured confounders may have introduced bias. Second, MR analysis assumes the genetic variants used as IVs are associated with VitD levels but not with other factors affecting rupture risk. Third, although we selected genetic variants strongly associated with VitD levels, they may have unexplained pleiotropic effects. Fourth, MR assumes a linear relationship between VitD levels and rupture, which may not hold if a threshold or nonlinear effect exists.
Additionally, other clinical factors, including age, diabetes, and multiple aneurysms, were significantly associated with aSAH in our observational study but were excluded from MR analysis due to limited genetic instruments and research focus. Moreover, our MR analysis utilized GWAS data from European populations, whereas our observational cohort was Chinese. Genetic differences in VitD metabolism, such as SLC24A5, CYP2R1, and GC variants between Europeans and Asians, may limit the generalizability of our MR findings [23, 24]. Dietary, lifestyle, and environmental factors, such as sunlight exposure and pollution, could also affect VitD‐related genotype expression.
Future studies using East Asian GWAS data are needed to validate these results in more genetically similar cohorts. Finally, not all relevant SNPs were included, which may have influenced the robustness of causal inference. A large‐scale prospective study with long‐term follow‐up is still needed to confirm our findings.
Source: Discussion Source: Discussion Source: Discussion Source: Discussion Source: Discussion Source: Discussion Source: DiscussionResults
Associations of Vitamin D With IA Rupture in the Observational Study
The baseline characteristics are presented in Table 1. A total of 499 patients with 619 IAs were included in this study, comprising 337 (67.5%) females and 162 (32.5%) males. The median age was 58.0 years [52.0, 64.0], and the median BMI was 24.8 kg/m2 [22.3, 27.3]. Among these patients, 184 (36.9%) had ruptured IAs, and 105 (21.0%) had multiple IAs.
Baseline information of enrolled patients before or after propensity score matching (PSM).
| Characteristic | Before PSM | After PSM | ||||||
|---|---|---|---|---|---|---|---|---|
| Overall (n = 499) | Unruptured (n = 315) | Ruptured (n = 184) | p | Overall (n = 224) | Unruptured (n = 112) | Ruptured (n = 112) | p | |
| Sex (Female) | 337 (67.5) | 215 (43.1) | 122 (24.4) | 0.654 a | 157 (70.1) | 80 (35.7) | 77 (34.4) | 0.662 a |
| Age (year) | 0.026 a | 0.774 a | ||||||
| ≤ 50 | 110 (22.0) | 58 (11.6) | 52 (10.4) | 43 (19.2) | 24 (10.7) | 19 (8.5) | ||
| 51 to 60 | 178 (35.7) | 110 (22.0) | 68 (13.6) | 85 (37.9) | 41 (18.3) | 44 (19.6) | ||
| 61 to 70 | 168 (33.7) | 118 (23.6) | 50 (10.0) | 78 (34.8) | 37 (16.5) | 41 (18.3) | ||
| > 70 | 43 (8.6) | 29 (5.8) | 14 (2.8) | 18 (8.0) | 9 (7.3) | 10 (8.1) | ||
| BMI | 0.699 a | 0.625 b | ||||||
| Underweight | 15 (3.0) | 9 (1.8) | 6 (1.2) | 7 (3.1) | 3 (1.3) | 4 (1.8) | ||
| Normal | 188 (37.9) | 114 (22.8) | 75 (15.0) | 89 (39.7) | 48 (21.4) | 41 (18.3) | ||
| Overweight | 204 (40.7) | 134 (26.9) | 69 (13.8) | 94 (42.0) | 47 (21.0) | 47 (21.0) | ||
| Obese | 92 (18.4) | 58 (11.6) | 34 (6.8) | 34 (15.2) | 14 (6.3) | 20 (8.9) | ||
| Smoking | 85 (17.0) | 50 (10.0) | 35 (7.0) | 0.367 a | 36 (16.1) | 19 (8.5) | 17 (7.6) | 0.716 a |
| Drinking | 53 (10.6) | 29 (5.8) | 24 (4.8) | 0.180 a | 25 (11.2) | 11 (4.9) | 14 (6.3) | 0.524 a |
| Diabetes | 70 (14.0) | 53 (10.6) | 17 (3.4) | 0.019 a | 20 (8.9) | 9 (4.0) | 11 (4.9) | 0.639 a |
| Hypertension | 291 (58.3) | 184 (36.9) | 107 (21.4) | 0.955 a | 132 (58.9) | 63 (28.1) | 69 (30.8) | 0.415 a |
| Hyperlipidemia | 142 (28.5) | 91 (18.2) | 51 (10.2) | 0.780 a | 56 (25.0) | 28 (12.5) | 28 (12.5) | 1.000 a |
| Blood test dates | 0.147 a | 0.885 a | ||||||
| Mar: May | 130 (26.1) | 81 (16.2) | 49 (9.8) | 60 (26.8) | 28 (12.5) | 32 (14.3) | ||
| Jun: Aug | 100 (20.0) | 57 (11.4) | 43 (8.6) | 45 (20.1) | 23 (10.3) | 22 (9.8) | ||
| Sep: Nov | 156 (31.3) | 96 (19.2) | 60 (12.0) | 73 (32.6) | 36 (16.1) | 37 (16.5) | ||
| Dec: Feb | 113 (22.6) | 81 (16.2) | 32 (6.4) | 46 (20.5) | 25 (11.2) | 21 (9.4) | ||
| Anterior circulating aneurysm | 477 (95.6) | 301 (60.3) | 176 (35.3) | 0.960 a | 216 (96.4) | 110 (49.1) | 106 (47.3) | 0.280 a |
| Multiple aneurysms | 105 (21.0) | 56 (11.2) | 49 (9.8) | 0.019 a | 45 (20.1) | 21 (9.4) | 24 (10.7) | 0.617 a |
| Vitamin D (ng/mL) | 17.0 [12.1 to 22.4] | 17.6 [12.6 to 22.6] | 16.3 [11.6 to 22.3] | 0.418 c | 16.0 [11.5 to 21.3] | 15.8 [11.7 to 21.4] | 16.2 [10.8 to 21.2] | 0.904 c |
| Vitamin D levels | 0.339 a | 0.947 a | ||||||
| Normal | 46 (9.2) | 26 (5.2) | 20 (4.0) | 22 (9.8) | 11 (4.9) | 11 (4.9) | ||
| Insufficient | 133 (26.7) | 90 (18.0) | 43 (8.6) | 48 (21.4) | 23 (10.3) | 25 (11.2) | ||
| Deficient | 320 (64.1) | 199 (39.9) | 121 (24.2) | 154 (68.8) | 78 (34.8) | 76 (33.9) | ||
Note: Variables are presented as number of patients (%) or median (interquartile range), as appropriate.
p‐values calculated using Pearson’s chi‐square test.
p‐values calculated using Fisher’s exact test.
p‐values calculated using Mann: Whitney U test.
In this study, diabetes was inversely associated with the risk of aSAH (p = 0.034; odds ratio [OR], 0.53; 95% confidence interval [CI], 0.290 to 0.954), whereas multiple aneurysms increased the risk of aSAH (p = 0.006; OR, 1.88; 95% CI, 1.197 to 2.940) in multivariate analysis (Table S1). The prevalence of diabetes was 16.8% (53/315) in patients with unruptured IAs, compared with 9.2% (17/184) in patients with ruptured IAs. Conversely, multiple aneurysms were more prevalent in ruptured IA patients (26.6%, 49/184) than in unruptured IA patients (17.8%, 56/315).
Age, sex, BMI, smoking, alcohol use, hypertension, hyperlipidemia, and aneurysm location (anterior circulation) showed no significant association with aSAH in multivariate analysis (p > 0.05).
Blood test dates, as presented in Table 1, did not differ significantly between patients with unruptured IA and those with ruptured IA (p = 0.147). The median VitD level was 17.0 ng/mL [12.1 to 22.4]. VitD levels were normal in 9.2% of patients, insufficient in 26.7%, and deficient in 64.1% of patients. No significant difference in VitD levels was observed between patients with unruptured IA and those with ruptured IA (p = 0.418). Furthermore, VitD deficiency showed no association with aSAH in this study (p = 0.339).
To control for confounders, PSM was performed in a 1:1 ratio based on sex, age, BMI, diabetes, hypertension, hyperlipidemia, smoking, alcohol consumption, blood test date, and presence of multiple aneurysms. A total of 112 patients with ruptured IA were successfully matched to 112 patients with unruptured IA. Following PSM, no significant differences were observed in the distribution of the aforementioned variables. VitD levels and deficiency remained unassociated with aSAH (p > 0.05).
Additionally, RCS regression analysis showed no significant nonlinear association between serum VitD levels and the risk of IA rupture (p = 0.739), nor a significant overall association (p = 0.865), further supporting the findings before and after PSM (Figure 2).
Restricted cubic spline regression model depicting the association between serum vitamin D (VitD) levels and the odds of ruptured intracranial aneurysms.
The Causal Associations Between Vitamin D Status and IA Rupture in MR Analysis
The IVW‐MR analysis revealed no causal relationship between VitD levels and the risks of aSAH (p = 0.966; OR, 1.00; 95% CI, 0.82 to 1.23) or unruptured IA (p = 0.74; OR, 0.95; 95% CI, 0.68 to 1.31) (Figure 3). These findings remained consistent across multiple MR methods, including weighted median, MR‐Egger, weighted mode, and simple mode approaches, and were replicated in an independent cohort (Table 2). Scatter plots and forest plots further confirmed these findings (Figures S1 and S2).
All SNPs used as instrumental variables (IVs) exhibited F‐statistics exceeding 20, indicating minimal weak instrument bias (Files S2: , S5). Sensitivity and heterogeneity analyses revealed no evidence of pleiotropy or heterogeneity (Table 3), supporting the robustness and reliability of the MR findings. Leave‐one‐out analysis further validated these results (Figure S3).
Forest plots for the causal associations of vitamin D (VitD) levels with ruptured intracranial aneurysms or unruptured intracranial aneurysms (UIAs) in discovery cohort.
The MR results of the replication cohort.
| Exposure | MR method | SNP | p | OR (95% CI) | |
|---|---|---|---|---|---|
| RIA | Vitamin D | MR Egger | 93 | 0.957 | 0.993 (0.777 to 1.270) |
| Weighted median | 93 | 0.417 | 0.891 (0.675 to 1.177) | ||
| Inverse variance weighted | 93 | 0.389 | 1.076 (0.911 to 1.270) | ||
| Simple mode | 93 | 0.745 | 0.907 (0.504 to 1.633) | ||
| Weighted mode | 93 | 0.948 | 1.007 (0.821 to 1.234) | ||
| UIA | Vitamin D | MR Egger | 93 | 0.410 | 0.874 (0.636 to 1.202) |
| Weighted median | 93 | 0.231 | 0.803 (0.561 to 1.150) | ||
| Inverse variance weighted | 93 | 0.371 | 1.103 (0.889 to 1.369) | ||
| Simple mode | 93 | 0.005 | 2.910 (1.413 to 5.990) | ||
| Weighted mode | 93 | 0.596 | 0.928 (0.706 to 1.221) |
Abbreviations: RIA, Ruptured intracranial aneurysms; UIA, Unruptured intracranial aneurysms.
Heterogeneity and pleiotropy tests of MR in discovery and replication cohort.
| Cohort | Test | Outcome | Exposure | Method | Effect size | p |
|---|---|---|---|---|---|---|
| Discovery | Heterogeneity | RIA | Vitamin D | Cochran’s Q test | 66.0 (QMR Egger) | 0.169 |
| 66.1 (QIVW) | 0.192 | |||||
| UIA | Vitamin D | Cochran’s Q test | 69.8 (QMR Egger) | 0.101 | ||
| 70.2 (QIVW) | 0.112 | |||||
| Pleiotropy | RIA | Vitamin D | MR‐Egger regression | 0.001 (Egger regression intercept) | 0.825 | |
| UIA | Vitamin D | MR‐Egger regression | −0.005 (Egger regression intercept) | 0.580 | ||
| Replication | Heterogeneity | RIA | Vitamin D | Cochran’s Q test | 110.8 (Q MR Egger) | 0.078 |
| 111.7 (Q IVW) | 0.080 | |||||
| UIA | Vitamin D | Cochran’s Q test | 88.6 (Q MR Egger) | 0.551 | ||
| 92.4 (Q IVW) | 0.469 | |||||
| Pleiotropy | RIA | Vitamin D | MR‐Egger regression | 0.003 (Egger regression intercept) | 0.393 | |
| UIA | Vitamin D | MR‐Egger regression | 0.009 (Egger regression intercept) | 0.055 |
Abbreviations: RIA, ruptured intracranial aneurysms; UIA, unruptured intracranial aneurysms.
Conclusion
Our study did not demonstrate an association or a causal relationship between VitD levels and aneurysmal subarachnoid hemorrhage (aSAH). The inverse associations observed in previous cohort studies may be attributable to confounding factors or reverse causation. Although VitD deficiency is common in patients with aSAH, supplementation may not prevent aneurysm rupture.
Source: ConclusionAuthors’ abstract
ABSTRACT
Background
Vitamin D (VitD) may protect arterial health, but its link to cerebral aneurysm rupture remains unclear. This study aims to investigate the correlation between serum VitD levels and the risk of aneurysmal subarachnoid hemorrhage (aSAH).
Methods
This retrospective study included patients with ruptured or unruptured aneurysms treated between 2018 and 2021. Univariate and multivariate analyses were conducted. Propensity score matching (PSM) analysis was conducted to balance confounding factors. Furthermore, the causal relationship between VitD and aSAH was investigated using a two‐sample Mendelian randomization (MR) analysis. Summary‐level data for the exposure (VitD) and outcome (aSAH) were obtained from public genome‐wide association study datasets. Multiple MR methods, including inverse variance weighted, MR‐Egger, weighted median, simple mode, and weighted mode, were utilized to assess causality.
Sensitivity analyses were conducted to evaluate the robustness and reliability of the causal estimates.
Results
A total of 499 patients with 619 intracranial aneurysms were included. Among them, 337 (68%) were female, 184 cases (36.9%) had ruptured aneurysms, and 105 (21.0%) had multiple aneurysms. VitD levels and VitD deficiency showed no association with aSAH in univariate or multivariate analysis before PSM (p > 0.05). After PSM (112 matched aSAH patients), VitD levels and VitD deficiency remained unassociated with the risk of aSAH (p = 0.947). MR analysis, including inverse variance weighted methods, found no causal link between VitD levels and aSAH (OR: 1.00; 95% CI: 0.82 to 1.23; p = 0.966).
Conclusion
This study found that serum VitD levels are neither associated with nor causally linked to aSAH. The inverse associations observed in previous studies may be attributed to confounding factors or reverse causation. A prospective, large‐scale study with long‐term follow‐up is warranted to validate these findings.
Read the full paper
Introduction
Intracranial aneurysms (IAs) are abnormal dilations of the cerebral arterial wall, posing a significant health risk due to their potential rupture, which can cause subarachnoid hemorrhage (SAH), a life‐threatening condition with high morbidity and mortality [1]. Incidental IAs are increasingly diagnosed owing to the widespread use of noninvasive neuroimaging. However, not all unruptured IAs rupture during a patient’s lifetime. Clinical factors such as hypertension, smoking, history of SAH, as well as aneurysm size and location, have been associated with IA growth and rupture [2, 3].
Tools such as the ELAPSS and PHASES scores, which incorporate these factors, have been developed to predict the risk of IA growth and rupture [4, 5]. Nevertheless, the literature lacks appropriate hematological indices for predicting aneurysm rupture.
While the pathological mechanisms underlying IA rupture remain incompletely understood, factors such as inflammation, oxidative stress, hemodynamic stress, and genetic predispositions are considered contributing factors [2, 3]. Abnormal blood flow and shear stress may compromise vascular integrity, while inflammation and oxidative stress could further damage the aneurysm wall, thereby increasing the rupture risk [2, 3, 6]. Vitamin D (VitD), a fat‐soluble steroid vitamin essential for calcium homeostasis and immune regulation, may confer vascular health benefits due to its anti‐inflammatory and antioxidative properties [7, 8].
Low levels of VitD are associated with elevated vascular inflammation, oxidative stress, endothelial dysfunction, and increased risk of cardiovascular and cerebrovascular events, including hypertension, atherosclerosis, myocardial infarction, and stroke [7, 8, 9]. VitD deficiency is also prevalent among IA patients [10, 11]. However, its relationship with IA rupture remains unclear. One study reported an inverse correlation between VitD levels and the risk of aneurysmal subarachnoid hemorrhage (aSAH) [12], whereas another found no association [11].
These discrepancies may arise from variations in sample size, study populations, methodologies, or unaccounted confounders. Additionally, VitD deficiency could result from aneurysm rupture, suggesting potential reverse causality. Thus, the specific role of VitD in aSAH remains controversial, and further research is required to determine whether VitD deficiency is associated with IA rupture.
Therefore, in this study, a large case: control study with propensity score matching (PSM) to balance confounding factors was conducted to investigate the potential association between VitD levels and aSAH. However, observational studies remain vulnerable to unmeasured confounders. To address this limitation, we also performed a two‐sample Mendelian randomization (MR) analysis to explore potential causal relationships. By integrating these methods, this study aims to provide high‐quality evidence regarding the relationship between VitD status and aSAH. The findings may improve our understanding of IA rupture mechanisms.
Materials and Methods
Observational Study Design
This retrospective case: control study comprised consecutive patients with IA treated at Qilu Hospital of Shandong University between March 2018 and May 2021. The study was approved by the Ethics Committee of Qilu Hospital of Shandong University, and written informed consent was obtained from all participants prior to enrollment. All IAs were confirmed using computed tomography (CT), magnetic resonance imaging (MRI), or digital subtraction angiography (DSA). SAH was confirmed via CT scans. Ruptured IAs were defined as cases presenting with aneurysmal SAH, regardless of the number of aneurysms present, while unruptured IAs were defined as patients with at least one aneurysm and no history of SAH.
Exclusion criteria were as follows: (1) recent administration of VitD supplements or glucocorticoids within 1 month prior to the study; (2) concomitant vascular diseases, such as arteriovenous malformations, moyamoya disease, or arteriovenous fistulas; (3) recurrent, fusiform, traumatic, bacterial, or dissecting aneurysms; (4) aneurysmal SAH occurring more than 72 h before admission; (5) SAH of unknown etiology or non‐aneurysmal SAH; and (6) age < 18 years. Ultimately, 499 patients with either unruptured or ruptured IAs were included in the study. The study flowchart is presented in Figure 1A.
The flow chart of the present study.
The patients’ demographic characteristics, comorbidities (including diabetes mellitus, hypertension, and hyperlipidemia), and aneurysm characteristics (such as location and number) were recorded. For patients with multiple aneurysms, the study evaluated the location (anterior or posterior circulation) of the primary aneurysm treated during hospitalization. Fasting venous blood samples were collected within 24 h of admission to measure 25‐hydroxyvitamin D levels and conduct lipid profile tests, including triglycerides (TG), total cholesterol (TC), low‐density lipoprotein cholesterol (LDL‐C), and high‐density lipoprotein cholesterol (HDL‐C). Body mass index (BMI) was categorized as underweight (≤ 18.5 kg/m2), normal weight (18.5 to 24 kg/m2), overweight (24 to 28 kg/m2), or obese (> 28 kg/m2).
Hypertension was defined as a documented history of the condition (systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg), irrespective of antihypertensive treatment. Diabetes mellitus was defined as patients receiving antidiabetic medication, a fasting glucose level ≥ 7 mmol/L from previous health assessments, or a hemoglobin A1c level ≥ 6.5%. Hyperlipidemia was defined as lipid‐lowering medication use or fasting plasma TC ≥ 6 mmol/L, TG ≥ 2 mmol/L, or LDL ≥ 3.5 mmol/L. Current smoking was defined as consumption of ≥ 5 cigarettes per day, and alcohol consumption as intake of ≥ 150 g per week.
VitD levels were categorized as deficient (< 20 ng/mL), insufficient (20 to 29 ng/mL), or normal (≥ 30 ng/mL).
Statistical analyses were performed using SPSS 25.0 (IBM Corp.). Continuous variables were expressed as mean ± standard deviation or median [interquartile range], depending on their distribution. The Shapiro: Wilk test was used to assess normality, with a p‐value < 0.05 indicating a deviation from normal distribution. Group comparisons were conducted using Student’s t‐test for normally distributed variables and the Mann: Whitney U test for non‐normally distributed variables. Categorical variables were expressed as frequencies and percentages and were compared using the χ 2 test or Fisher’s exact test.
Factors with p < 0.1 in univariate analysis were included in binary logistic regression analysis using an enter process to identify independent predictors. PSM was used to balance confounding factors between patients with and without aneurysmal SAH. A one‐to‐one PSM analysis using the nearest neighbor method with a caliper value of 0.01 was conducted to adjust for imbalanced factors. A restricted cubic spline (RCS) regression model was employed to examine potential linear or nonlinear associations between serum VitD levels and the risk of aSAH.
Four knots were positioned at the 5th, 35th, 65th, and 95th percentiles of the VitD distribution, with the median serving as the reference point. The RCS model was fitted using the “rms” package (version 6.8 to 2), and spline plots were generated using the “plotRCS” package (version 0.1.5) in R software (version 4.4.1). Statistical significance was defined as a two‐tailed p‐value < 0.05.
MR Analysis
A two‐sample MR analysis was performed to investigate the causal relationship between VitD levels and aSAH. Figure 1B shows the analysis flowchart.
Genetic instruments for VitD levels were obtained from genome‐wide association studies (GWAS) datasets available through the IEU OpenGWAS project (https://gwas.mrcieu.ac.uk/), a public resource for investigating genetic associations with complex traits. Summary‐level GWAS data on IAs were derived from two large cohorts: the discovery cohort, conducted by the International Stroke Genetics Consortium (ISGC) with individuals of European ancestry, and the replication cohort from FinnGen. The discovery cohort included 5140 ruptured cases, 2070 unruptured cases, and 71,934 controls, while the replication cohort had 6236 ruptured cases, 3014 unruptured cases, and 412,105 controls.
Additional details regarding the data sources used in this study are provided in File S1.
From these GWAS datasets, single nucleotide polymorphisms (SNPs) significantly associated with VitD levels (p < 5 × 10−8) were selected as candidate instrumental variables (IVs). Linkage disequilibrium (LD) analysis was conducted to ensure SNP independence, applying a threshold of LD‐r 2 < 0.001 and a clumping distance threshold of > 10,000 kb. The F‐statistic [F = R 2(N‐2)/(1‐R 2)] was calculated to evaluate IV strength, with SNPs failing to meet the criterion of F ≥ 10 excluded to mitigate weak instrument bias, thereby maintaining the robustness of the analysis.
The inverse variance‐weighted (IVW) method was employed as the primary approach for causal estimation, providing reliable results in the absence of heterogeneity or horizontal pleiotropy. Supplementary methods: including MR‐Egger regression, weighted median, simple mode, and weighted mode: were used to ensure robustness under varying conditions. Heterogeneity was assessed using Cochran’s Q test, and if heterogeneity was detected (p < 0.05), adjusted results were reanalyzed [13]. Horizontal pleiotropy was evaluated using the MR‐Egger intercept method, with p > 0.05 indicating no pleiotropy and supporting MR reliability [14].
The leave‐one‐out (LOO) method was applied to systematically exclude each SNP to assess result stability. All analyses were conducted using R software (version 4.4.1) with the “TwoSampleMR” package [15], with statistical significance defined as p < 0.05.
Results
Associations of Vitamin D With IA Rupture in the Observational Study
The baseline characteristics are presented in Table 1. A total of 499 patients with 619 IAs were included in this study, comprising 337 (67.5%) females and 162 (32.5%) males. The median age was 58.0 years [52.0, 64.0], and the median BMI was 24.8 kg/m2 [22.3, 27.3]. Among these patients, 184 (36.9%) had ruptured IAs, and 105 (21.0%) had multiple IAs.
Baseline information of enrolled patients before or after propensity score matching (PSM).
| Characteristic | Before PSM | After PSM | ||||||
|---|---|---|---|---|---|---|---|---|
| Overall (n = 499) | Unruptured (n = 315) | Ruptured (n = 184) | p | Overall (n = 224) | Unruptured (n = 112) | Ruptured (n = 112) | p | |
| Sex (Female) | 337 (67.5) | 215 (43.1) | 122 (24.4) | 0.654 a | 157 (70.1) | 80 (35.7) | 77 (34.4) | 0.662 a |
| Age (year) | 0.026 a | 0.774 a | ||||||
| ≤ 50 | 110 (22.0) | 58 (11.6) | 52 (10.4) | 43 (19.2) | 24 (10.7) | 19 (8.5) | ||
| 51 to 60 | 178 (35.7) | 110 (22.0) | 68 (13.6) | 85 (37.9) | 41 (18.3) | 44 (19.6) | ||
| 61 to 70 | 168 (33.7) | 118 (23.6) | 50 (10.0) | 78 (34.8) | 37 (16.5) | 41 (18.3) | ||
| > 70 | 43 (8.6) | 29 (5.8) | 14 (2.8) | 18 (8.0) | 9 (7.3) | 10 (8.1) | ||
| BMI | 0.699 a | 0.625 b | ||||||
| Underweight | 15 (3.0) | 9 (1.8) | 6 (1.2) | 7 (3.1) | 3 (1.3) | 4 (1.8) | ||
| Normal | 188 (37.9) | 114 (22.8) | 75 (15.0) | 89 (39.7) | 48 (21.4) | 41 (18.3) | ||
| Overweight | 204 (40.7) | 134 (26.9) | 69 (13.8) | 94 (42.0) | 47 (21.0) | 47 (21.0) | ||
| Obese | 92 (18.4) | 58 (11.6) | 34 (6.8) | 34 (15.2) | 14 (6.3) | 20 (8.9) | ||
| Smoking | 85 (17.0) | 50 (10.0) | 35 (7.0) | 0.367 a | 36 (16.1) | 19 (8.5) | 17 (7.6) | 0.716 a |
| Drinking | 53 (10.6) | 29 (5.8) | 24 (4.8) | 0.180 a | 25 (11.2) | 11 (4.9) | 14 (6.3) | 0.524 a |
| Diabetes | 70 (14.0) | 53 (10.6) | 17 (3.4) | 0.019 a | 20 (8.9) | 9 (4.0) | 11 (4.9) | 0.639 a |
| Hypertension | 291 (58.3) | 184 (36.9) | 107 (21.4) | 0.955 a | 132 (58.9) | 63 (28.1) | 69 (30.8) | 0.415 a |
| Hyperlipidemia | 142 (28.5) | 91 (18.2) | 51 (10.2) | 0.780 a | 56 (25.0) | 28 (12.5) | 28 (12.5) | 1.000 a |
| Blood test dates | 0.147 a | 0.885 a | ||||||
| Mar: May | 130 (26.1) | 81 (16.2) | 49 (9.8) | 60 (26.8) | 28 (12.5) | 32 (14.3) | ||
| Jun: Aug | 100 (20.0) | 57 (11.4) | 43 (8.6) | 45 (20.1) | 23 (10.3) | 22 (9.8) | ||
| Sep: Nov | 156 (31.3) | 96 (19.2) | 60 (12.0) | 73 (32.6) | 36 (16.1) | 37 (16.5) | ||
| Dec: Feb | 113 (22.6) | 81 (16.2) | 32 (6.4) | 46 (20.5) | 25 (11.2) | 21 (9.4) | ||
| Anterior circulating aneurysm | 477 (95.6) | 301 (60.3) | 176 (35.3) | 0.960 a | 216 (96.4) | 110 (49.1) | 106 (47.3) | 0.280 a |
| Multiple aneurysms | 105 (21.0) | 56 (11.2) | 49 (9.8) | 0.019 a | 45 (20.1) | 21 (9.4) | 24 (10.7) | 0.617 a |
| Vitamin D (ng/mL) | 17.0 [12.1 to 22.4] | 17.6 [12.6 to 22.6] | 16.3 [11.6 to 22.3] | 0.418 c | 16.0 [11.5 to 21.3] | 15.8 [11.7 to 21.4] | 16.2 [10.8 to 21.2] | 0.904 c |
| Vitamin D levels | 0.339 a | 0.947 a | ||||||
| Normal | 46 (9.2) | 26 (5.2) | 20 (4.0) | 22 (9.8) | 11 (4.9) | 11 (4.9) | ||
| Insufficient | 133 (26.7) | 90 (18.0) | 43 (8.6) | 48 (21.4) | 23 (10.3) | 25 (11.2) | ||
| Deficient | 320 (64.1) | 199 (39.9) | 121 (24.2) | 154 (68.8) | 78 (34.8) | 76 (33.9) | ||
Note: Variables are presented as number of patients (%) or median (interquartile range), as appropriate.
p‐values calculated using Pearson’s chi‐square test.
p‐values calculated using Fisher’s exact test.
p‐values calculated using Mann: Whitney U test.
In this study, diabetes was inversely associated with the risk of aSAH (p = 0.034; odds ratio [OR], 0.53; 95% confidence interval [CI], 0.290 to 0.954), whereas multiple aneurysms increased the risk of aSAH (p = 0.006; OR, 1.88; 95% CI, 1.197 to 2.940) in multivariate analysis (Table S1). The prevalence of diabetes was 16.8% (53/315) in patients with unruptured IAs, compared with 9.2% (17/184) in patients with ruptured IAs. Conversely, multiple aneurysms were more prevalent in ruptured IA patients (26.6%, 49/184) than in unruptured IA patients (17.8%, 56/315).
Age, sex, BMI, smoking, alcohol use, hypertension, hyperlipidemia, and aneurysm location (anterior circulation) showed no significant association with aSAH in multivariate analysis (p > 0.05).
Blood test dates, as presented in Table 1, did not differ significantly between patients with unruptured IA and those with ruptured IA (p = 0.147). The median VitD level was 17.0 ng/mL [12.1 to 22.4]. VitD levels were normal in 9.2% of patients, insufficient in 26.7%, and deficient in 64.1% of patients. No significant difference in VitD levels was observed between patients with unruptured IA and those with ruptured IA (p = 0.418). Furthermore, VitD deficiency showed no association with aSAH in this study (p = 0.339).
To control for confounders, PSM was performed in a 1:1 ratio based on sex, age, BMI, diabetes, hypertension, hyperlipidemia, smoking, alcohol consumption, blood test date, and presence of multiple aneurysms. A total of 112 patients with ruptured IA were successfully matched to 112 patients with unruptured IA. Following PSM, no significant differences were observed in the distribution of the aforementioned variables. VitD levels and deficiency remained unassociated with aSAH (p > 0.05).
Additionally, RCS regression analysis showed no significant nonlinear association between serum VitD levels and the risk of IA rupture (p = 0.739), nor a significant overall association (p = 0.865), further supporting the findings before and after PSM (Figure 2).
Restricted cubic spline regression model depicting the association between serum vitamin D (VitD) levels and the odds of ruptured intracranial aneurysms.
The Causal Associations Between Vitamin D Status and IA Rupture in MR Analysis
The IVW‐MR analysis revealed no causal relationship between VitD levels and the risks of aSAH (p = 0.966; OR, 1.00; 95% CI, 0.82 to 1.23) or unruptured IA (p = 0.74; OR, 0.95; 95% CI, 0.68 to 1.31) (Figure 3). These findings remained consistent across multiple MR methods, including weighted median, MR‐Egger, weighted mode, and simple mode approaches, and were replicated in an independent cohort (Table 2). Scatter plots and forest plots further confirmed these findings (Figures S1 and S2).
All SNPs used as instrumental variables (IVs) exhibited F‐statistics exceeding 20, indicating minimal weak instrument bias (Files S2: , S5). Sensitivity and heterogeneity analyses revealed no evidence of pleiotropy or heterogeneity (Table 3), supporting the robustness and reliability of the MR findings. Leave‐one‐out analysis further validated these results (Figure S3).
Forest plots for the causal associations of vitamin D (VitD) levels with ruptured intracranial aneurysms or unruptured intracranial aneurysms (UIAs) in discovery cohort.
The MR results of the replication cohort.
| Exposure | MR method | SNP | p | OR (95% CI) | |
|---|---|---|---|---|---|
| RIA | Vitamin D | MR Egger | 93 | 0.957 | 0.993 (0.777 to 1.270) |
| Weighted median | 93 | 0.417 | 0.891 (0.675 to 1.177) | ||
| Inverse variance weighted | 93 | 0.389 | 1.076 (0.911 to 1.270) | ||
| Simple mode | 93 | 0.745 | 0.907 (0.504 to 1.633) | ||
| Weighted mode | 93 | 0.948 | 1.007 (0.821 to 1.234) | ||
| UIA | Vitamin D | MR Egger | 93 | 0.410 | 0.874 (0.636 to 1.202) |
| Weighted median | 93 | 0.231 | 0.803 (0.561 to 1.150) | ||
| Inverse variance weighted | 93 | 0.371 | 1.103 (0.889 to 1.369) | ||
| Simple mode | 93 | 0.005 | 2.910 (1.413 to 5.990) | ||
| Weighted mode | 93 | 0.596 | 0.928 (0.706 to 1.221) |
Abbreviations: RIA, Ruptured intracranial aneurysms; UIA, Unruptured intracranial aneurysms.
Heterogeneity and pleiotropy tests of MR in discovery and replication cohort.
| Cohort | Test | Outcome | Exposure | Method | Effect size | p |
|---|---|---|---|---|---|---|
| Discovery | Heterogeneity | RIA | Vitamin D | Cochran’s Q test | 66.0 (QMR Egger) | 0.169 |
| 66.1 (QIVW) | 0.192 | |||||
| UIA | Vitamin D | Cochran’s Q test | 69.8 (QMR Egger) | 0.101 | ||
| 70.2 (QIVW) | 0.112 | |||||
| Pleiotropy | RIA | Vitamin D | MR‐Egger regression | 0.001 (Egger regression intercept) | 0.825 | |
| UIA | Vitamin D | MR‐Egger regression | −0.005 (Egger regression intercept) | 0.580 | ||
| Replication | Heterogeneity | RIA | Vitamin D | Cochran’s Q test | 110.8 (Q MR Egger) | 0.078 |
| 111.7 (Q IVW) | 0.080 | |||||
| UIA | Vitamin D | Cochran’s Q test | 88.6 (Q MR Egger) | 0.551 | ||
| 92.4 (Q IVW) | 0.469 | |||||
| Pleiotropy | RIA | Vitamin D | MR‐Egger regression | 0.003 (Egger regression intercept) | 0.393 | |
| UIA | Vitamin D | MR‐Egger regression | 0.009 (Egger regression intercept) | 0.055 |
Abbreviations: RIA, ruptured intracranial aneurysms; UIA, unruptured intracranial aneurysms.
Discussion
The relationship between VitD status and the risk of aSAH has received increasing attention due to accumulating evidence demonstrating the high prevalence of VitD deficiency in the population and the potential protective role of VitD in cardiovascular and cerebrovascular health [7, 8, 9, 16]. However, our study, which employed a large case: control design with PSM and two‐sample MR analysis, did not identify any association or causal relationship between VitD status and the risk of IA rupture.
VitD has a potential protective role in maintaining arterial health [17]. VitD receptors are present in vascular smooth muscle and endothelial cells. VitD deficiency is common in patients with IAs, even in those with aSAH and high sun exposure [18]. Although VitD might protect against IAs and reduce the risk of aSAH, our findings did not support this hypothesis, as VitD levels were not inversely associated with aSAH. A previous cross‐sectional study found higher VitD levels in patients with unruptured IAs than in those with ruptured IAs, with VitD independently associated with rupture risk [12].
However, that study had a small sample size and lacked PSM analysis. Our use of PSM balanced baseline characteristics between ruptured and unruptured IA groups, minimizing confounders such as age, sex, and comorbidities. Despite this, we found no significant difference in VitD levels between patients with ruptured or unruptured IA, and VitD deficiency was not associated with aSAH. Our results were consistent with a previous study [10], which found no difference in VitD levels between ruptured and unruptured IA patients, despite higher rates of hypovitaminosis D in IA patients requiring treatment compared to patients without IA.
However, our observational study may still be affected by uncontrolled confounders, such as lifestyle, dietary habits, sunlight exposure, and diseases or medications influencing VitD metabolism, which could potentially impact serum VitD levels and our findings. Future studies incorporating detailed lifestyle and environmental data are warranted to better account for these influences.
To address this limitation, we performed MR analysis, which utilizes genetic variants as instrumental variables to infer causal relationships. This analysis revealed no causal association between genetically determined VitD levels and IA rupture, consistent with prior MR studies utilizing distinct GWAS cohorts that also found no causal effect of VitD on IA rupture. We further used a restricted cubic spline regression model to evaluate potential nonlinear associations between serum VitD levels and IA rupture risk. The analysis revealed no significant linear or nonlinear relationship, thereby reinforcing our findings.
These results suggest that previously reported associations may be attributable to residual confounding or unmeasured biases. Although lower VitD levels were frequently observed in aneurysm patients, a recent meta‐analysis of VitD intervention trials failed to demonstrate a significant reduction in cardiovascular events with VitD supplementation, suggesting that lower VitD levels may represent an epiphenomenon rather than a causative factor [19]. One recent study reported VitD deficiency could promote IA rupture in a murine model [20]. However, all SAH murine models have inherent limitations, and the specific animal model used in that study may not fully replicate human aneurysm rupture mechanisms.
Therefore, these findings cannot be directly extrapolated to humans.
The biological mechanisms through which VitD influences vascular health may not directly relate to the pathophysiology of IA rupture. Although VitD may contribute to general vascular maintenance, the specific processes leading to aneurysm formation and rupture: such as hemodynamic stress, vascular wall degeneration, and inflammatory infiltration: may be less affected by VitD levels. Additionally, the VitD threshold for vascular protection may be lower than deficiency levels, meaning even “low” VitD levels could suffice for vascular integrity. However, studies indicate that VitD has anti‐inflammatory effects [21], and cerebrovascular inflammation is a major pathological factor for aneurysm rupture.
Low VitD levels are also associated with risk factors for aSAH, such as hypertension, hyperlipidemia, and hyperhomocysteinemia [22]. Therefore, our findings should be interpreted with caution. An alternative explanation is that the relationship between VitD and IA rupture is more complex than a linear association. VitD may directly or indirectly influence measurable and unmeasurable factors that contribute to the pathological progression of aneurysms. Its vascular effects might also be more pronounced in specific subgroups, such as individuals with certain comorbidities or genetic predispositions, and these effects may have been attenuated in our overall study population.
Although our study did not identify a significant association between VitD levels and IA rupture, this finding does not rule out VitD’s potential role in other aspects of aneurysm biology, such as formation or progression. Future studies should investigate these areas through longitudinal study designs or by focusing on subgroups more susceptible to the effects of VitD. Further research should also explore interactions between VitD and other factors, such as genetic variation, environmental exposures, or comorbidities, which may modulate IA rupture risk.
For instance, VitD might exert protective effects only in individuals with specific genetic profiles or under particular environmental conditions. Elucidating these interactions could provide deeper insights into VitD’s role in vascular pathophysiology.
The strengths of our study include the utilization of PSM and MR analysis to mitigate confounding and establish causal inferences. PSM effectively balances control groups and reduces selection bias in observational studies, while MR reliably evaluates causality by employing genetic variants as instrumental variables for VitD levels, thereby minimizing confounding and reverse causation. However, the MR analysis design differs from our observational analysis. While the observational study compared VitD levels in patients with ruptured and unruptured IA, the MR analysis assessed genetically predicted VitD levels in relation to unruptured IA and aSAH risk versus general population controls.
Nevertheless, these complementary methods provide both real‐world and genetically informed insights into the potential role of VitD in the pathophysiology of IA.
However, our study has several limitations. First, while PSM balanced observed baseline characteristics, unmeasured confounders may have introduced bias. Second, MR analysis assumes the genetic variants used as IVs are associated with VitD levels but not with other factors affecting rupture risk. Third, although we selected genetic variants strongly associated with VitD levels, they may have unexplained pleiotropic effects. Fourth, MR assumes a linear relationship between VitD levels and rupture, which may not hold if a threshold or nonlinear effect exists.
Additionally, other clinical factors, including age, diabetes, and multiple aneurysms, were significantly associated with aSAH in our observational study but were excluded from MR analysis due to limited genetic instruments and research focus. Moreover, our MR analysis utilized GWAS data from European populations, whereas our observational cohort was Chinese. Genetic differences in VitD metabolism, such as SLC24A5, CYP2R1, and GC variants between Europeans and Asians, may limit the generalizability of our MR findings [23, 24]. Dietary, lifestyle, and environmental factors, such as sunlight exposure and pollution, could also affect VitD‐related genotype expression.
Future studies using East Asian GWAS data are needed to validate these results in more genetically similar cohorts. Finally, not all relevant SNPs were included, which may have influenced the robustness of causal inference. A large‐scale prospective study with long‐term follow‐up is still needed to confirm our findings.
Conclusion
Our study did not demonstrate an association or a causal relationship between VitD levels and aneurysmal subarachnoid hemorrhage (aSAH). The inverse associations observed in previous cohort studies may be attributable to confounding factors or reverse causation. Although VitD deficiency is common in patients with aSAH, supplementation may not prevent aneurysm rupture.
Ethics Statement
The study was approved by the ethics committee of Qilu Hospital of Shandong University, and written informed consent was obtained from all the patients before enrollment.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Figure S1: Scatter plots of the relationship between vitamin D levels and ruptured or unruptured intracranial aneurysms. Scatter plots illustrate the associations between vitamin D levels and ruptured or unruptured intracranial aneurysms in the discovery cohort (A and C) and the replication cohort (B and D). MR, Mendelian randomization; SNP, single nucleotide polymorphism.
Figure S2: Causal effects of individual SNPs in MR analyses. Forest plots show the estimated MR effect sizes of individual SNPs on ruptured or unruptured intracranial aneurysms by vitamin D levels in the discovery cohort (A and C) and the replication cohort (B and D). MR, Mendelian randomization; SNP, single nucleotide polymorphism.
Figure S3: Leave‐one‐out sensitivity analysis of MR results. Leave‐one‐out sensitivity analyses are shown for the discovery cohort (A and C) and the replication cohort (B and D). MR, Mendelian randomization.
File S1: cns70617‐sup‐0002‐FileS1.xlsx.
File S2: cns70617‐sup‐0003‐FileS2.xlsx.
File S3: cns70617‐sup‐0004‐FileS3.xlsx.
File S4: cns70617‐sup‐0005‐FileS4.xlsx.
File S5: cns70617‐sup‐0006‐FileS5.xlsx.
Table S1: Binary logistic regression between patients with ruptured intracranial aneurysms and patients with unruptured intracranial aneurysms.
Data Availability Statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
References
Publication and source details
Original title: Serum Vitamin D Levels Were Not Associated With the Risk of Aneurysmal Subarachnoid Hemorrhage: A Large Cohort Study With Propensity Score Matching and Mendelian Randomization Analysis
Haoran Qiu; Kai Chen; Yang Yu; Ziyin Song; Jingzheng Liu; Lvyin Luo; Xinlong Ma; Zhaoyang Yuan; Maogui Li; Jianfeng Zhuang; Mingxiang Zhang; Wandong Su; Yunyan Wang; Donghai Wang; Weiying Zhong
CNS Neuroscience & Therapeutics · 2025
* Correspondence:Weiying Zhong (zhongweiying123@163.com)
© 2025 The Author(s). CNS Neuroscience & Therapeutics published by John Wiley & Sons Ltd.https://creativecommons.org/licenses/by/4.0/This is an open access article under the terms of the http://creativecommons.org/licenses/by/4.0/ License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Natural Science Foundation of Shandong Province10.13039/501100007129 ZR2022MH131
Repository JATS snapshot; version as supplied
Source retrieved: 2026-11-11T01:03:52.184999+00:00
DOI: 10.1111/cns.70617 · PMC12481830