Vitamin D, Hypertension, and Ischemic Stroke in 116 655 Individuals From the General PopulationNovelty and Significance
A Genetic Study
Observational studies indicate that low concentrations of plasma 25-hydroxyvitamin D (25(OH)D) are associated with high blood pressure, hypertension, and ischemic stroke. However, whether these associations are causal remain unknown. A total of 116 655 white individuals of Danish descent from the general population were genotyped for genetic variants in DHCR7 and CYP2R1 affecting plasma 25(OH)D concentrations; 35 517 had plasma 25(OH)D measurements. Primary outcomes were blood pressure, hypertension, and ischemic stroke. Median follow-up for incident ischemic stroke was 9.3 years (range, 1 day–33.6 years). DHCR7/CYP2R1 allele score was as expected associated with lower 25(OH)D concentration (F=328 and R2=1.0%). A genetically determined 10 nmol/L lower 25(OH)D concentration was associated with a 0.68 (95% confidence interval [CI], 0.20–1.17) mm Hg higher systolic blood pressure and a 0.36 (95% CI, 0.08–0.63) mm Hg higher diastolic blood pressure with corresponding observational estimates of 0.58 (95% CI, 0.50–0.68) and 0.40 (95% CI, 0.35–0.45) mm Hg. The odds ratio for hypertension was 1.02 (95% CI, 0.97–1.08) for a genetically determined 10 nmol/L lower 25(OH)D with a corresponding observational odds ratio of 1.06 (95% CI, 1.05–1.07). The odds ratio for ischemic stroke was 0.98 (95% CI, 0.86–1.13) for a genetically determined 10 nmol/L decrease in 25(OH)D with a corresponding observational odds ratio of 1.03 (95% CI, 1.01–1.05). Genetic and observational low 25(OH)D concentrations were associated with higher blood pressure, as well as with hypertension consistent with causal relationships. Because observational but not genetic low 25(OH)D concentration was associated with ischemic stroke, and as the CIs overlapped, we can neither support nor exclude a causal relationship.
See Editorial Commentary, pp 496–498
Observational studies indicate that low concentrations of plasma 25-hydroxyvitamin D (25(OH)D), usually used to asses vitamin D status, are associated with higher blood pressure, hypertension, and ischemic stroke.1–4 In addition, a Mendelian randomization study has shown an increased risk of hypertension with genetically low 25(OH)D.3 However, randomized studies have shown minor to no effects of vitamin D supplementation on lowering of blood pressure and cardiovascular disease risk.4–6 Furthermore, genetic studies show that some risk factors for ischemic stroke, such as obesity and an atherogenic lipid profile, may be causally associated with low concentrations of 25(OH)D.7–9 Thus, it is unclear whether low 25(OH)D is a cause of high blood pressure, hypertension, and ischemic stroke or whether the associations are largely a result of confounding and reverse causation.
The use of genetic variants in Mendelian randomization studies allows for analyses less susceptible to confounding and free of reverse causation because the random assortment of genetic variants that occurs during gamete formation secures an equal distribution of confounding factors among different genotypes and genotypes are not affected by outcome10,11; thus, genetic variants in DHCR7 and CYP2R1 that specifically lower 25(OH)D concentrations provide an instrument for assessing the potential consequences of lifelong low 25(OH)D concentrations on blood pressure, hypertension, and ischemic stroke, largely free of confounding and free of reverse causation.
We tested the hypothesis that genetically low 25(OH)D concentrations are associated with high blood pressure, hypertension, and ischemic stroke (Figure 1). First, in observational analyses, we tested the association of 25(OH)D concentrations with blood pressure, hypertension, and ischemic stroke (Figure 1A, arrow 1); second and third, whether the selected genotypes were associated with plasma 25(OH)D concentrations and with blood pressure, hypertension, and ischemic stroke (Figure 1A, arrows 2 and 3); and fourth, whether the selected genotypes were associated with blood pressure, hypertension, and ischemic stroke consistent with their effect on 25(OH)D concentrations by using instrumental variable analysis (Figure 1A, arrow 4).
The CCHS (Copenhagen City Heart Study) was initiated in 1976 to 1978 with follow-up examinations after 5 (1981–1983), 15 (1991–1994), and 25 years (2001–2003).12 Individuals aged 20 to 100 years were randomly invited from the national Danish Central Person Register to reflect the Danish general population. In observational analyses, we included 9896 individuals with plasma 25(OH)D measurements from the 1981 to 1983 examination and in genetic analyses, 9740 individuals with all genotypes from the 1991 to 1994 and 2001 to 2003 examinations. Of these, 5458 individuals had both 25(OH)D measurements and genotype data available.
The CGPS (Copenhagen General Population Study) was initiated in 2003 with ongoing enrollment and with individuals recruited as for to the CCHS.7,8 In observational analyses, we included 25 621 individuals with plasma 25(OH)D measurements and in genetic analyses, 106 915 individuals with all genotypes. Of these, 25 518 had both 25(OH)D measurements and genotype data available.
The studies were approved by institutional review boards and Danish ethical committees, and individuals provided written informed consent. No individuals appeared in >1 study, and all were white of Danish descent. A flowchart depicting how we arrived at our sample sizes is presented in Figure 1B.
Plasma 25(OH)D Measurements
We used the DiaSorin Liaison 25(OH)D TOTAL assay blinded to outcome and genotype data. CGPS plasma samples were collected in 2004 to 2005 (n=12 501; stored at −80°C for ≈5 years) and in 2009 to 2011 (n=13 120; measured on fresh samples) while CCHS plasma samples were collected in 1981 to 1983 (n=9896; stored at −20°C for ≈26 years); all samples were collected on the day of examination. Assay precision was tested daily while assay accuracy was tested using an external quality control program (DEQAS). The interassay coefficient of variance was 10% for a low level control (≈40 nmol/L) and 8% for a high level control (≈135 nmol/L). Samples for measurement were consecutive individuals for the time periods mentioned for CGPS and all available plasma samples from the CCHS 1981 to 1983 examination.
Genotyping using TaqMan assays was conducted blinded to 25(OH)D concentration and outcome data. Genotypes were selected among those having the strongest, largest association with 25(OH)D concentration in genome-wide association studies13,14; genetic variants around DHCR7 (rs7944926 and rs11234027) and CYP2R1 (rs10741657 and rs12794714) were specifically chosen because they are expected to influence 25(OH)D concentration through synthesis of 25(OH)D. We deliberately did not include polymorphisms in the vitamin D–binding protein because these do not associate predictably with 25(OH)D’s biological activity.15 Genotypes were verified by sequencing of 32 randomly selected samples in the 2 cohorts. Call rates for the genotypes were >99% after 2 reruns. For DHCR7, CYP2R1, and DHCR7/CYP2R1 allele scores, weighted allele scores were constructed by multiplying each variant allele with its effect on plasma 25(OH)D concentration adjusted for the effect of the other variant in each gene, for example, the effect of DHCR7 rs7944926 on plasma 25(OH)D was adjusted for DHCR7 rs11234027, because these 2 genotypes are correlated. Weighted allele score were used for all allele score and instrumental variable analyses. Furthermore, for sensitivity analyses, allele scores were created by simply counting all alleles across the 4 genotypes instead of the more complex weighted scores.16
Confounders were chosen based on the known important risk factors for ischemic stroke and their possible association with plasma 25(OH)D.1 Individuals reported on smoking status, daily tobacco consumption, alcohol consumption, intensity of leisure time physical activity, income, diabetes mellitus, and occurrence of stroke in parents, and all information was reviewed together with an investigator on the day of attendance. Cumulative tobacco consumption was calculated in pack-years, where 1 pack-year was 20 cigarettes or equivalent smoked daily for 1 year. Body mass index was measured weight (kg) divided by measured height (m) squared on the day of examination. Furthermore, baseline atrial fibrillation was determined by registry diagnoses. Standard hospital assays were used to measure total and high-density lipoprotein (HDL) cholesterol; non-HDL cholesterol was total minus HDL cholesterol. Last, we adjusted for kidney function using estimated glomerular filtration rate (CKD-EPI equation [Chronic Kidney Disease Epidemiology Collaboration])17 because kidney function affects both blood pressure and 25(OH)D levels.
Brachial systolic and diastolic blood pressures on the left arm (mm Hg) were measured on the day of examination by trained technicians either using a London School of Hygiene sphygmomanometer or an automatic Digital Blood Pressure Monitor (Kivex) as a single measurement on the left arm after 5 minutes of rest and with the subject in the sitting position.18 Hypertension was defined as self-reported use of antihypertensive medication as systolic blood pressure ≥140 mm Hg or as diastolic blood pressure ≥90 mm Hg. Severe hypertension was defined as self-reported use of antihypertensive medication as systolic blood pressure ≥160 mm Hg or as diastolic blood pressure ≥100 mm Hg. Blood pressure was adjusted for use of antihypertensives by adding 10 and 5 mm Hg to systolic and diastolic blood pressures, respectively.19
Diagnosis of cerebrovascular disease, including ischemic and hemorrhagic strokes, was collected from 1976 until November 2014 by reviewing hospital admissions with diagnoses entered in the national Danish Patient Registry and causes of death entered in the national Danish Causes of Death Registry.2,12,20 For individuals with registered cerebrovascular disease, records from general practitioners and hospital were requested, and the diagnosis of ischemic stroke was validated by 2 independent doctors with special interest in stroke according to World Health Organization criteria and classified into subgroups using clinical description, computed tomography or MRI scan, spinal fluid examination, autopsy, or surgical description.20 Median follow-up for incident ischemic stroke was 9.3 years (range, 1 day–33.6 years). Case-fatality rate, defined as death within 30 days of an ischemic stroke event, was 5.8%.
We used Stata/S.E. 13.1. χ2 tests evaluated Hardy–Weinberg equilibrium. More than 99% of observations were present for the included variables. We imputed missing data by using multivariable chained imputation (mi impute chained) with fully conditional specification; however, results were similar without using imputation. All analyses were adjusted for age, sex, and study as a minimum; in addition, observational analyses were adjusted for smoking status, cumulative tobacco consumption, alcohol consumption, leisure time physical activity, body mass index, income, diabetes mellitus, ratio of non-HDL to HDL cholesterol, stroke in parents, atrial fibrillation, and month and year of blood sample. Hypertension was not included in the ischemic stroke model because it could be a mediator from low 25(OH)D to ischemic stroke. All analyses with blood pressure and hypertension as outcome were cross-sectional, whereas the analyses with ischemic stroke were prospective when using measured plasma 25(OH)D and the allele scores as exposures. To maximize power the instrumental variable estimates for ischemic stroke, all registered cases were used.
First, we tested whether plasma 25(OH)D concentrations were observationally associated with blood pressure, hypertension, and ischemic stroke (Figure 1A, arrow 1). For these analyses, we used multiple linear regression, logistic regression, and Cox regression models with age as time scale, respectively. The 25(OH)D concentrations were categorized as deficient (<25 nmol/L), insufficient (25–49 nmol/L), and sufficient (≥50 nmol/L)21 and were also modeled using nonlinear terms in the regression models. Specifically, we used restricted cubic splines with 3 knots for nonlinear associations.22
Second, we used Cuzick nonparametric trend test to assess trend across genotypes and allele scores of 25(OH)D concentrations. We assessed the strengths of genotypes and allele scores as instruments by using the F statistic and R2 as a measure of variation explained by genotypes and allele scores (Figure 1A, arrow 2).11
Third, we examined associations of DHCR7/CYP2R1 allele score with blood pressure, hypertension, and ischemic stroke using the same models as in observational analyses; however, these analyses were only adjusted for age, sex, month and year of blood sample, and study because allele scores were randomly distributed across potential confounders (Figure 1A, arrow 3).
Fourth, we calculated instrumental variable estimates per 10 nmol/L lower 25(OH)D by using the Wald-type ratio estimator, which involves taking the ratio of the outcome allele score coefficient to the exposure allele score coefficient and for odds ratios exponentiating to express the estimate as an odds ratio (Figure 1A, arrow 4).11,23 We used the delta method to derive SEs of Wald-type instrumental variable log odds ratios.24 For comparison, we derived observational estimates per 10 nmol/L lower 25(OH)D by using multiple linear regression for blood pressure and logistic regression models for hypertension and ischemic stroke adjusted for age, sex, and study. Additional details can be found in the online-only Data Supplement.
Plasma 25(OH)D concentration was associated with all major risk factors for ischemic stroke except for occurrence of stroke in parents (Table). In contrast, the DHCR7/CYP2R1 allele score was not associated with these potential confounders (Table S1 in the online-only Data Supplement), illustrating that the allele score can be used as a largely unconfounded instrument to assess the association of genetically low 25(OH)D with blood pressure, hypertension, and ischemic stroke. DHCR7 and CYP2R1 genotypes were not in linkage disequilibrium (R2=0%), implying that genetic variants in the 2 genes were completely unrelated. Within each gene, the variants each explained 49% of the variation in the other. However, there were no linkage disequilibrium with other genetic variants associated with hypertension and stroke on chromosome 11 identified through genome-wide association studies (Figure S1). The characteristics of individuals included in the observational and genetic studies were comparable (Table S2).
25(OH)D, Blood Pressure, Hypertension, and Ischemic Stroke: Observational Estimates
We tested the association of 25(OH)D concentrations with blood pressure, hypertension, and incident ischemic stroke using in cubic spline models that indicated an almost linear increase in blood pressure and risk of hypertension and ischemic stroke with decreasing 25(OH)D concentrations <50 nmol/L (Figure 2).
Systolic blood pressures was 2.56 (95% confidence interval [CI], 1.85–3.27) mm Hg higher for individuals with 25(OH)D of <25 versus ≥50 nmol/L; the corresponding difference in diastolic blood pressure was 1.88 (95% CI, 1.47–2.29) mm Hg (Figure S2). The multivariable-adjusted odds ratio for hypertension was 1.28 (95% CI, 1.18–1.38) for individuals with 25(OH)D of <25 versus ≥50 nmol/L. The multivariable-adjusted hazard ratio for incident ischemic stroke was 1.23 (95% CI, 1.06–1.42) for individuals with 25(OH)D of <25 versus ≥50 nmol/L.
Genotypes and Plasma 25(OH)D
The combined unweighted DHCR7/CYP2R1 allele score was associated with 8.4 (95% CI, 7.4–9.5) nmol/L lower 25(OH)D concentration for 6 to 8 versus 0 to 1 variant alleles; the F-value was 328 and R2 was 1.0% (Figure 3). For each variant allele, 25(OH)D was lowered by 1.9 (95% CI, 1.7–2.1) nmol/L for the unweighted allele score. Each increase in weighted allele score corresponded to a 1 nmol/L lowering of 25(OH)D concentration. The association of individual genotypes with 25(OH)D showed similar results though with less power (Figure S3).
Allele Score and Blood Pressure, Hypertension, and Ischemic Stroke: Genetic Estimates
Systolic blood pressure was 0.07 (95% CI, 0.02–0.12) mm Hg higher per 1-unit increase in the weighted DHCR7/CYP2R1 allele score; the corresponding estimate for diastolic blood pressure was 0.04 (95% CI, 0.01–0.06) mm Hg (Figure 4). The odds ratio for hypertension was 1.002 (95% CI, 0.997–1.007) per one increase in the weighted DHCR7/CYP2R1 allele score. The corresponding hazard ratio for incident ischemic stroke was 0.997 (95% CI, 0.980–1.013).
25(OH)D, Blood Pressure, Hypertension, and Ischemic Stroke: Instrumental Variable and Observational Estimates
A genetically determined 10 nmol/L lower 25(OH)D concentration was associated with a 0.68 (95% CI, 0.20–1.17) mm Hg higher systolic blood pressure and a 0.36 (95% CI, 0.08–0.63) mm Hg higher diastolic blood pressure (Figure 5). Corresponding observational estimates were 0.59 (95% CI, 0.50–0.68) and 0.40 (95% CI, 0.35–0.45) mm Hg, respectively. The odds ratio for hypertension was 1.02 (95% CI, 0.97–1.08) for genetically determined 10 nmol/L lower 25(OH)D concentration. The corresponding observational odds ratio was 1.06 (95% CI, 1.05–1.07). The odds ratio for ischemic stroke was 0.98 (95% CI, 0.86–1.13) for genetically determined 10 nmol/L lower 25(OH)D concentration. The corresponding observational odds ratio was 1.03 (95% CI, 1.01–1.05).
Using individual genotypes or unweighted alleles scores, the results were similar to those using the weighted allele score for all outcomes (Figures S4–S6). Because the 2 genotypes within DHCR7 and CYP2R1 were correlated, supplementary instrumental variable analyses using only one genotype from each gene were performed, and results were similar to the main analyses for associations with 25(OH)D and for blood pressure, hypertension, and ischemic stroke (Figures S7–S9). Furthermore, the instrumental variable for systolic and diastolic blood after exclusion of those on antihypertensive medication showed similar results to those presented in Figure 5 albeit with reduced power (Figure S10). Finally, the genetic analyses were restricted to those with a 25(OH)D measurement, and the results were statistically comparable to results presented in Figure 5 albeit with reduced power (ie, broad CIs; Figure S11).
In addition, we compared the estimates for difference in blood pressure with each genetic variant in our study with publically available genome-wide association data (International Consortium for Blood Pressure25,26); for all 4 genetic variants, the estimates were comparable with the estimates from International Consortium for Blood Pressure (Figure S12). Likewise, we compared estimates from our study with a previous Mendelian randomization Study on 25(OH)D concentrations and blood pressure and hypertension3; the results were again similar (Figure S13).
Also, we tested the association of 25(OH)D concentration with severe hypertension defined as systolic/diastolic blood pressure >160/100 mm Hg or use antihypertensive medication (Figure S14). The odds ratio for severe hypertension was 1.10 (95% CI, 1.04–1.17) for a genetically determined 10 nmol/L lower 25(OH)D concentration. The corresponding observational odds ratio was 1.04 (95% CI, 1.02–1.05). Last, we investigated our blood pressure measurements for last digit preference; there was a tendency for rounding to 0 and to a lesser degree 5 (Figure S15).
Using a Mendelian randomization approach in 116 655 individuals from the general population, we found that genetic and observational low 25(OH)D concentrations were associated with high blood pressure and hypertension compatible with causal relationships. Because observational but not genetic low 25(OH)D concentration was associated with ischemic stroke, and as the CIs overlapped, we can neither support nor exclude a causal relationship.
Biological mechanisms proposed to link low vitamin D concentrations with blood pressure include effects on the renin–angiotensin system and arterial wall thickness or stiffness. Some animal and human studies suggest that the active form of vitamin D, 1,25-dihydroxyvitamin D, may suppress renin secretion,27,28 whereas other studies show little effect of 1,25-dihydroxyvitamin D on renin secretion.29 Likewise, results from randomized intervention trials investigating the effects of vitamin D supplementation on arterial stiffness have been conflicting, some supporting a reduction in arterial stiffness while other studies show no effect.30 Also, blood pressure is affected by plasma concentrations of parathyroid hormone and calcium that are suppressed and increased, respectively, by vitamin D supplementation, indicating that vitamin D may have indirect effects on blood pressure through other molecules.31 Thus, plausible mechanisms have been suggested that link low 25(OH)D with increased blood pressure and hypertension, but the evidence supporting these mechanisms is conflicting.
The results from the present study are at odds with recent randomized intervention trials showing no effects of vitamin D supplementation on blood pressure in normotensive and hypertensive individuals.5 This could be explained by the nonlinear association of 25(OH)D concentration with blood pressure observed in the present study, where the inverse association was primarily present for 25(OH)D <50 nmol/L with an average estimated effect of ≈0.5 mm Hg increase in blood pressure per 10 nmol/L decrease in 25(OH)D. Thus, little effect is expected in individuals with 25(OH)D >50 nmol/L, and the sample size required to show a 1-mm Hg change in systolic blood pressure with 80% power in a randomized intervention trials would be ≈3000 individuals (based on data from the DAYLIGHT trial [The Vitamin D Therapy in Individuals at High Risk of Hypertension Trial]32). Furthermore, although the implication of a positive finding in a Mendelian randomization study is the presence of causality, it should be remembered that the setting is different from a randomized intervention trial, that is, we investigated lifelong exposure to low 25(OH)D and not short-term intervention with vitamin D. Nonetheless, a previous Mendelian randomization study has also shown an effect of genetically low 25(OH)D concentration on high blood pressure and hypertension, indicating that these findings are robust.3
Although a modest genetic effect on blood pressure could be shown, a clear genetic effect of low vitamin D could not be seen on risk of ischemic stroke in the present study. Given the small effect size on blood pressure and failure to demonstrate causal associations of low 25(OH)D with cardiovascular risk factors and outcomes in previous studies,7–9,33 this is perhaps not surprising. In principle, this could be a power issue because the present study had 80% power to show odds ratios of ≥1.5; however, given the present results and the risk of adverse events,4 supplementation with vitamin D for prevention of ischemic stroke does not seem like a viable option for general clinical practice.
Strengths of our study include that we had enough statistical power to detect an association of 25(OH)D lowering genotypes with blood pressure. Furthermore, we did not detect any violations of the assumptions underlying Mendelian randomization as far as they could be tested, and our instruments used for instrumental variable analyses were strong with an F-value of 328. In addition, because individuals were included at random and consecutively from the general population, both for genetic and 25(OH)D analyses, the potential for selection bias is minimal.
The Mendelian randomization approach assumes absence of genetic pleiotropy and linkage disequilibrium with other genetic variants associated with the outcome for the genetic variants used as instruments. However, as shown previously, there is no evidence of genetic pleiotropy,7,34 and the variants affecting 25(OH)D concentration are not in linkage disequilibrium with other genetic variants associated with blood pressure, atrial fibrillation, or ischemic stroke in genome-wide association studies. Furthermore, 25(OH)D concentrations are known to vary with sun exposure and skin color, and we only studied white Danes, thus potentially limiting the generalizability of the results to other geographical regions. However, population homogeneity does eliminate population admixture as a potential confounder in our study. Ideally, we would have preferred to have 25(OH)D measurements in all included individuals; however, although our study is one of the largest cohorts with plasma 25(OH)D measurements, the cost is too high for measurement in all participants with genotypes. Furthermore, Mendelian randomization approaches are not dependent on complete measurement of phenotype; rather, several approaches advocate use of subsets or independent samples with phenotype or genotype measurements, respectively, to reduce bias, save cost, and maximize power instead of restricting to analyses to samples with complete measurements, which could introduce bias.35–38 Other potential limitations are that we only measured blood pressure once and that we adjusted for use of antihypertensive medication by adding 10 and 5 mm Hg to systolic and diastolic blood pressures. This adds measurement noise to our data decreasing power as the first read in a blood pressure measurement may yield higher blood pressures on average and as the correction may be more or less precise for each individual. However, our sensitivity analyses indicate that results were similar when excluding those using antihypertensive medication and using a higher cut off for defining hypertension yielded stronger evidence for a causal effect. Thus, these potential biases are unlikely to explain any positive findings. Last, given the small effects sizes and somewhat limited power for ischemic stroke, we can neither support nor exclude that low 25(OH)D leads to ischemic stroke.
In summary, genetic and observational low 25(OH)D concentrations were associated with higher blood pressure, as well as with hypertension consistent with causal relationships. Because observational but not genetic low 25(OH)D concentration was associated ischemic stroke, and as the CIs overlapped, we can neither support nor exclude a causal relationship. Thus, our study indicates the need for larger genetic studies and larger randomized intervention trials investigating the effects of vitamin D on ischemic stroke. However, given the modest effects of vitamin D status on blood pressure, the data do not support a clear-cut recommendation for vitamin D supplementation for reducing blood pressure to prevent cardiovascular complications, such as ischemic stroke.
We thank the staff and participants of the Copenhagen General Population Study for their important contributions to our study.
Sources of Funding
This work was supported by Herlev and Gentofte Hospital, Copenhagen University Hospital. The funding source had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.
The online-only Data Supplement is available with this article at http://hyper.ahajournals.org/lookup/suppl/doi:10.1161/HYPERTENSIONAHA.117.09411/-/DC1.
- Received March 19, 2017.
- Revision received April 11, 2017.
- Accepted June 19, 2017.
- © 2017 American Heart Association, Inc.
- Vimaleswaran KS,
- Cavadino A,
- Berry DJ,
- et al
- Theodoratou E,
- Tzoulaki I,
- Zgaga L,
- Ioannidis JP
- Elamin MB,
- Abu Elnour NO,
- Elamin KB,
- Fatourechi MM,
- Alkatib AA,
- Almandoz JP,
- Liu H,
- Lane MA,
- Mullan RJ,
- Hazem A,
- Erwin PJ,
- Hensrud DD,
- Murad MH,
- Montori VM
- Ooi EM,
- Afzal S,
- Nordestgaard BG
- Vimaleswaran KS,
- Berry DJ,
- Lu C,
- et al
- Timpson NJ,
- Harbord R,
- Davey Smith G,
- Zacho J,
- Tybjaerg-Hansen A,
- Nordestgaard BG
- Wain LV,
- Verwoert GC,
- O’Reilly PF,
- et al
- Schroten NF,
- Ruifrok WP,
- Kleijn L,
- Dokter MM,
- Silljé HH,
- Lambers Heerspink HJ,
- Bakker SJ,
- Kema IP,
- van Gilst WH,
- van Veldhuisen DJ,
- Hillege HL,
- de Boer RA
- Vaidya A,
- Brown JM,
- Williams JS
- Arora P,
- Song Y,
- Dusek J,
- et al
- Brøndum-Jacobsen P,
- Benn M,
- Afzal S,
- Nordestgaard BG
- Afzal S,
- Brøndum-Jacobsen P,
- Bojesen SE,
- Nordestgaard BG
- Hartwig FP,
- Davies NM,
- Hemani G,
- Davey Smith G
Novelty and Significance
What Is New?
We used a Mendelian randomization design to test whether the association of low 25-hydroxyvitamin D with high blood pressure, hypertension, and ischemic stroke is causal or the result of confounding and reverse causation. One previous large scale Mendelian randomization has investigated the association with high blood pressure and hypertension while the genetic association with ischemic stroke has not been investigated in this setting before.
What Is Relevant?
Vitamin D supplementation is easily administered, and if low 25-hydroxyvitamin D is causally associated with high blood pressure, hypertension, or ischemic stroke, it could have wide implications for public health strategies.
Given the present results with a causal but modest effect of low vitamin D on high blood pressure, a clear-cut recommendation for supplementation with vitamin D for prevention of hypertension cannot be given. Further studies are required to investigate potential minor effects if any on risk of ischemic stroke.