Special Article – Vitamin D
Austin J Nutri Food Sci. 2019; 7(3): 1119.
Circulating 25-Hydroxyvitamin D levels and Risk of Incident Stroke: An Updated Meta-Analysis
Jong-Myon Bae*
Department of Preventive Medicine, Jeju National University College of Medicine, Korea
*Corresponding author: Jong-Myon Bae, Department of Preventive Medicine, Jeju National University College of Medicine, 102 Jejudaehak-ro, Jeju-si, 63243, Jeju Province, Korea
Received: May 06, 2019; Accepted: June 18, 2019; Published: June 25, 2019
Abstract
A recent systematic review that selected 19 relevant articles that were published up until September 30, 2017 showed that vitamin D deficiency was associated with Ischemic Stroke (IS) but not Hemorrhagic Stroke (HS). However, heterogeneity would be introduced when comparing the lowest and highest categories of vitamin D. The aim of this article was to conduct an Updated Meta-Analysis (UMA) that involved searching for relevant articles published up until March 31, 2019. An interval collapsing method was applied for information extraction to decrease heterogeneity among studies. Additional articles were selected from cited lists from 19 selected articles using citation discovery tools. The random effect model was applied if the I-squared value exceeded 50%. A funnel plot and Egger’s test were used to detect publication bias. After the addition of five new studies, the relative risks [and their 95% confidence intervals] (and I-squared value) were 1.52 [1.33-1.74] (0.0%) for IS and 2.44 [1.34-4.46] (69.7%) for HS. This UMA supported the hypothesis that serum vitamin D deficiency was associated with an increased risk of HS as well as IS. Diverse public health programs targeting vitamin D deficiency are needed for higher-risk groups, such as the older population.
Keywords: Vitamin D; Stroke; Risk factors; Systematic Review; Metaanalysis
Abbreviations
25(OH)D: 25-Hydroxyvitamin D; CDT: Citation Discovery Tools; CI: Confidence Interval; FES: First-Ever Stroke; HLM: Highest Versus Lowest Method; HS: Hemorrhagic Stroke; ICM: Interval Collapsing Method; IS: Ischemic Stroke (IS); logRR: Logarithm Relative Risk; OS: Overall Stroke; RR: Relative Risk; SElogRR: Standard Error of Logarithm Relative Risk; sRR: Summary Relative Risk
Introduction
As stoke is a leading case of mortality and disability globally [1], the economic burden is substantial [2,3]. Although hypertension, diabetes mellitus, obesity, and stroke are well known as important risk factors of stroke, the exploration of unknown risk factors is still needed [2,4].
Several studies reported that the incidence of First-Ever Stroke (FES) is higher in winter and spring [5]. Similar to tuberculosis [6] or suicide [7] that show seasonal variation of occurrence, the hypothetical association between vitamin D deficiency and risk of FES has been suggested [8-10]. Zhou et al. [4] conducted a quantitative systematic review of 19 relevant articles [11-29] published up to 30 September 2017, and concluded that vitamin D level was associated with Ischemic Stroke (IS), but not Hemorrhagic Stroke (HS).
However, the following two issues were identified with the study by Zhou et al [4]. First, they did not specify the method by which vitamin D was measured from blood sampling or intake amounts. Among the 19 selected articles, Kojima et al. [15] and Ford et al. [22] evaluated the vitamin D level of subjects using a food frequency questionnaire and supplement intake data, respectively. The remaining articles assessed vitamin D levels by measuring serum 25-Hydroxyvitamin D [25(OH) D]. Second, Michos et al. [16] having the outcome as mortality was selected for meta-analysis, even though the aim of Zhou et al. [4] was to verify the association between vitamin D level and the ‘incidence’ of stroke. Thus, it is necessary to perform an Updated Meta-Analysis (UMA) to clarify the results in Zhou et al [4]. The aim of this UMA was to evaluate the hypothesis that lower level of circulating 25(OH) D are associated with an increased risk of stroke.
Materials and Methods
As Zhou et al. [4] selected relevant articles that were published up to September 30, 2017, it is necessary to add relevant studies that have been published up until 31 March 2019. A search list was created through the Citation Discovery Tools (CDT) of “cited by” provided by PubMed [30] from the 19 articles selected by Zhou et al [4]. The inclusion and exclusion criteria were as same as the study by Zhou et al [4]. In other words, the selection criteria were analytic epidemiological studies that measured the circulating 25(OH)D levels of cohort participants and identified the risk of HS as well as IS and Overall Stroke (OS).
Instead of the ‘highest versus lowest’ method (HLM) used by Zhou et al. [4], an ‘Interval Collapsing Method’ (ICM) was used to extract information from each selected article to make full use of the information in the selected articles [31,32]. The Logarithm Relative Risk (log RR) and the standard error of log RR (SE log R) for each article were calculated from the extracted Relative Risk (RR) and 95% confidence intervals (CI).
The heterogeneity of articles was assessed by the I-squared value (%). A random effect model was used when the I-squared value exceeded 50%, whereas if this value was below 50% a fixed effect model was used [33]. Subgroup analyses were conducted by study design, such as cohort and case-control. Publication bias was evaluated by funnel plot and Egger’s test. If a publication bias was confirmed, sensitivity analysis was performed with limiting SElogRR. The level of statistical significance was set to 0.05.
Results
A total of 359 studies were retrieved from the 19 studies selected by Zhou et al. [4] using PubMed’s CDT. Five studies were additionally selected [34-38]. Zhang et al. [34] and Manouchehri et al. [35] were published after 30 September 2017. With the addition of 16 studies [11-14,17-21,23-29], 21 studies were finally selected for meta-analysis (Table 1). These included 14 cohort studies [11-14,17,19-21,23- 26,28,34] and 7 case-control studies [18,27,29,35-38].
Reference number
First Author
Year
Design
Types of stroke
No. of Subjects
logRR
SElogRR
Study or Nation
11
Marniemi
2005
COS
OS
755
-0.07
0.23
Finland
12
Anderson
2010
COS
OS
41 504
0.41
0.13
IHC
13
Bolland
2010
COS
OS
1471
0.34
0.29
New Zealand
14
Drechsler
2010
COS
OS
1108
0.9
0.36
4D
17
Schierbeck
2012
COS
OS
2 016
0.52
0.22
DOPS
18
Sun
2012
CCS
IS
928
0.31
0.13
NHS
19
Kuhn
2013
COS
OS
3 115
-0.05
0.14
EPIC Germany
20
Perna
2013
COS
OS
9949
0.22
0.1
ESTHER
21
Skaaby
2013
COS
OS
9146
-0.12
0.1
Monica 10 & Inter 99
23
Schneider
2015
COS
OS
12 158
0.11
0.06
ARIC
24
Judd
2016
COS
OS
1 547
0.43
0.15
REGARDS
IS
0.37
0.16
HS
0.49
0.24
25
Zittermann
2016
COS
OS
154
0.89
0.41
Germany
IS
0.86
0.57
HS
0.65
0.54
26
Afzal
2017
COS
OS
116 655
0.13
0.07
CCHS
27
Alfieri
2017
CCS
IS
286
1.5
1.29
Brazil
28
Leung
2017
COS
OS
3458
0.38
0.11
Hong Kong
IS
0.36
0.12
HS
0.46
0.23
29
Tan
2017
CCS
OS
404
1.47
1.08
China
IS
1.42
1.04
HS
1.55
1.14
34
Zhang
2019
COS
OS
4 808
-0.09
0.1
WHI-OS
35
Manouchehri
2017
CCS
HS
150
1.97
0.39
Iran
36
Afshari
2015
CCS
IS
72
1.26
0.64
Iran
37
Gupta
2014
CCS
IS
143
0.3
0.29
India
38
Chaudhuri
2014
CCS
IS
500
0.72
0.19
India
*CCS: Case-Control Study; CI: Confidence Interval; COS: Cohort Study; logRR: Logarithm Relative Risk; HS: Hemorrhagic Stroke; IS: Ischemic Stroke; OS: Overall Stroke: SElogRR: Standard Error Of Logarithm Relative Risk
Table 1: Summary table of the extracted information from 21 selected studies*.
From the 21 studies, the summary RR (sRR) [95% CI] (I-squared value, %) of OS, IS, and HS were 1.36 [1.19-1.55] (74.3%), 1.52 [1.33- 1.74] (0.0%), and 2.44 [1.34-4.46] (69.7%), respectively (Table 2) (Figure 1). When subgroup analyses were conducted, the results from three cohort studies and two case-control studies showed statistical significance regarding the risk of HS.
Figure 1: Forest plot for estimating the summary effect size (ES) in all 21 selected studies.
Overall stroke
Ischemic stroke
Hemorrhagic stroke
All selected
1.36 [1.19-1.55] (74.3) {21}
1.52 [1.33-1.74] (0.0) {9}
2.44 [1.34-4.46] (69.7) {5}
Cohort
1.21 [1.07-1.36] (67.3) {14}
1.46 [1.22-1.76] (0.0) {3}
1.63 [1.20-2.22] (0.0) {3}
Case-control
2.32 [1.44-3.73] (71.4) {7}
1.59 [1.31-1.93] (25.1) {6}
6.87 [3.35-14.0] (0.0) {2}
Table 2: Summary relative risks [95% confidence intervals] (I squared value, %) in {number} of selected articles by types of stroke.
Egger’s test on the 21 studies suggested publication bias (P=0.003) (Table 3). When the test was performed to the 15 studies with SElogRR< 0.3, the publication bias disappeared (P=0.129) (Figure 2), and the sOR of OS remained statistically significant.
Figure 2: Funnel plot for 15 studies having standard error of log relative risk
(s.e. of logRR) less than 0.3 (P-value of Egger’s test = 0.129).
Overall stroke
Ischemic stroke
Hemorrhagic stroke
All selected
1.36 [1.19-1.55] (74.3) {21}
1.52 [1.33-1.74] (0.0) {9}
2.44 [1.34-4.46] (69.7) {5}
Cohort
1.21 [1.07-1.36] (67.3) {14}
1.46 [1.22-1.76] (0.0) {3}
1.63 [1.20-2.22] (0.0) {3}
Case-control
2.32 [1.44-3.73] (71.4) {7}
1.59 [1.31-1.93] (25.1) {6}
6.87 [3.35-14.0] (0.0) {2}
Table 3: Summary relative risks [95% confidence intervals] (I squared value, %) in {number} of selected articles from restriction of standard error of log relative risk (SElogRR) and their P-value of Egger’s test.
Discussion
Taken together, the results indicated that a lower level of circulating 25(OH)D was associated with a significant increase of 1.36-fold for the risk of OS, 1.52-fold for the risk of IS, and 2.44-fold for the risk of HS. Statistical significance was maintained in subgroup analysis conducted according to study design. In particular, this UMA showed that circulating vitamin D levels were associated with HS through adding Manouchehri et al. [35] and using ICM [31,32]. Zhou et al. [4] did not show the statistically significant association between vitamin D and HS risk.
Based on the findings, this UMA has two advantages. First, five studies were added using PubMed’s CDT, three [36-38] of which were published before 30 September 2017. In other words, they should have been selected in the analysis of Zhou et al [4]. This suggests that adding new relevant studies using CDT would be an efficient and valid methodology to conduct an UMA [30,39-41]. Second, ICM was employed to make full use of the suggested information, which was consistent with the original purpose of the meta-analysis [42]. It is necessary to consider ICM for meta-analysis of nutritional epidemiological studies that categorize according to the overall distribution rather than the absolute criteria [31], because Zhou et al. [4] mentioned that a limitation was the heterogeneity introduced by using HLM.
The major limitations and suggestion of this UMA are as follows. First, author did not evaluate the quality of selected articles using the Newcastle-Ottawa Scale (NOS) or Grading of recommendation, assessment, development and evaluation. Instead, author did conduct subgroup analyses by study design for observational studies in nutritional epidemiology. The reason was based on the suggestion by Bae JM [43], which concluded that ‘it is more reasonable to control for quality level by performing subgroup analysis according to study design rather than by using high quality based on the NOS quality assessment tool. “Second, publication biases was detected in selected studies for OS and IS, but not HS, but this was removed by restricting studies that had a SElogRR of below 0.3. The relationship between hypovitaminosis D and the risk of OS and IS was significant. Further analytical epidemiological studies for HS risk are needed because of the lack of research on HS compared to IS.
Conclusion
Despite above limitations, this UMA provided that a lower level of circulating vitamin D was associated with the risk of HS, as well as IS and OS. Thus, higher levels of circulating vitamin D might be a protective factor for HS, as well as IS.
References
- GBD 2016 Stroke Collaborators. Global, regional, and national burden of stroke, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol. 2019; 18: 439-458.
- Kim JY, Kang KS, Kang JH, Koo JS, Kim DH, Kim BJ, et al. Executive summary of stroke statistics in Korea 2018: a report from the Epidemiology Research Council of the Korean Stroke Society. J Stroke. 2019; 21: 42-59.
- Kim HJ, Kim YA, Seo HY, Kim EJ, Yoon SJ, Oh IH, et al. The economic burden of stroke in 2010 in Korea. J Korean Med Assoc. 2012; 55: 1226- 1236.
- Zhou R, Wang M, Huang H, Li W, Hu Y, Wu T, et al. Lower Vitamin D Status Is Associated with an Increased Risk of Ischemic Stroke: A Systematic Review and Meta-Analysis. Nutrients. 2018; 10: 277.
- Palm F, Dos Santos M, Urbanek C, Greulich M, Zimmer K, Safer A, et al. Stroke seasonality associations with subtype, etiology and laboratory results in the Ludwigshafen Stroke Study (LuSSt). Eur J Epidemiol. 2013; 28: 373- 381.
- Kim EH, Bae JM. Vitamin D supplementation as a control program against latent tuberculosis infection in Korean high school students. Epidemiol Health. 2018; 40: e2018035.
- Spedding S. Vitamin D and depression: A systematic review and metaanalysis comparing studies with and without biological flaws. Nutrients. 2014; 44: 31-35.
- Narasimhan S, Balasubramanian P. Role of Vitamin D in the Outcome of Ischemic Stroke- A Randomized Controlled Trial. J ClinDiagn Res. 2017; 11: CC06-CC10.
- Majeed F. Low levels of Vitamin D an emerging risk for cardiovascular diseases: A review. Int J Health Sci (Qassim). 2017; 11: 71-76.
- Tang Z, Li M, Zhang X, Hou W. Dietary flavonoid intake and the risk of stroke: a dose-response meta-analysis of prospective cohort studies. BMJ Open. 2016; 6(6): e008680.
- Marniemi J, Alanen E, Impivaara O, Seppänen R, Hakala P, Rajala T, et al. Dietary and serum vitamins and minerals as predictors of myocardial infarction and stroke in elderly subjects. Nutr Metab Cardiovasc Dis. 2005; 15: 188-197.
- Anderson JL, May HT, Horne BD, Bair TL, Hall NL, Carlquist JF, et al. Relation of vitamin D deficiency to cardiovascular risk factors, disease status, and incident events in a general healthcare population. Am J Cardiol. 2010; 106: 963-968.
- Bolland MJ, Bacon CJ, Horne AM, Mason BH, Ames RW, Wang TK, et al. Vitamin D insufficiency and health outcomes over 5 y in older women. Am J ClinNutr. 2010; 91: 82-89.
- Drechsler C, Pilz S, Obermayer-Pietsch B, Verduijn M, Tomaschitz A, Krane V, et al. Vitamin D deficiency is associated with sudden cardiac death, combined cardiovascular events, and mortality in haemodialysis patients. Eur Heart J. 2010; 31: 2253-2261.
- Skaaby T, Husemoen LL, Pisinger C, Jørgensen T, Thuesen BH, Fenger M, et al. Vitamin D status and incident cardiovascular disease and all-cause mortality: a general population study. Endocrine. 2013; 43: 618-625.
- Ford JA, MacLennan GS, Avenell A, Bolland M, Grey A, Witham M, et al. Cardiovascular disease and vitamin D supplementation: trial analysis, systematic review, and meta-analysis. Am J ClinNutr. 2014; 100: 746-755.
- Schneider AL, Lutsey PL, Selvin E, Mosley TH, Sharrett AR, Carson KA, et al. Vitamin D, vitamin D binding protein gene polymorphisms, race and risk of incident stroke: the Atherosclerosis Risk in Communities (ARIC) study. Eur J Neurol. 2015; 22: 1220-1227.
- Judd SE, Morgan CJ, Panwar B, Howard VJ, Wadley VG, Jenny NS, et al. Vitamin D deficiency and incident stroke risk in community-living black and white adults. Int J Stroke. 2016; 11: 93-102.
- Zittermann A, Morshuis M, Kuhn J, Pilz S, Ernst JB, Oezpeker C, et al. Vitamin D metabolites and fibroblast growth factor-23 in patients with left ventricular assist device implants: association with stroke and mortality risk. Eur J Nutr. 2016; 55: 305-313.
- Afzal S, Nordestgaard BG. Vitamin D, Hypertension, and Ischemic Stroke in 116 655 Individuals from the General Population: A Genetic Study. Hypertension. 2017.
- Alfieri DF, Lehmann MF, Oliveira SR, Flauzino T, Delongui F, de Araújo MC, et al. Vitamin D deficiency is associated with acute ischemic stroke, C-reactive protein, and short-term outcome. Metab Brain Dis. 2017; 32: 493-502.
- Leung RY, Han Y, Sing CW, Cheung BM, Wong IC, Tan KC, et al. Serum 25-hydroxyvitamin D and the risk of stroke in Hong Kong Chinese. ThrombHaemost. 2017; 117: 158-163.
- Tan LM, Wang L, Chen JJ, Li H, Luo WB. Diagnostic performance of bone metabolic indexes for the detection of stroke. Saudi Med J. 2017; 38: 30-35.
- Bae JM, Kim EH. Citation Discovery Tools for Conducting Adaptive Metaanalyses to Update Systematic Reviews. J Prev Med Public Health. 2016; 49: 129-133.
- Bae JM. Comparison of methods of extracting information for meta-analysis of observational studies in nutritional epidemiology. Epidemiol Health. 2016; 38: e2016003.
- Bae JM. Reinterpretation of the results of a pooled analysis of dietary carotenoid intake and breast cancer risk by using the interval collapsing method. Epidemiol Health. 2016; 38: e2016024.
- Harris RJ, Bradburn MJ, Deeks JJ, Harborad RM, Altman DG, Sterne JAC. Fixed- and random-effects meta-analysis. Stata J. 2008; 8: 3-28.
- Zhang X, Tu W, Manson JE, Tinker L, Liu S, Cauley JA, et al. Racial/Ethnic Differences in 25-Hydroxy Vitamin D and Parathyroid Hormone Levels and Cardiovascular Disease Risk Among Postmenopausal Women. J Am Heart Assoc. 2019; 8: e011021.
- Manouchehri N, Vakil-Asadollahi M, Zandifar A, Rasmani F, Saadatnia M. Vitamin D Status in Small Vessel and Large Vessel Ischemic Stroke Patients: A Case-control Study. Adv Biomed Res. 2017; 6: 146.
- Afshari L, Amani R, Soltani F, Haghighizadeh MH, Afsharmanesh MR. The relation between serum Vitamin D levels and body antioxidant status in ischemic stroke patients: A case-control study. Adv Biomed Res. 2015; 4: 213.
- Gupta A, Prabhakar S, Modi M, Bhadada SK, Lal V, Khurana D. Vitamin D status and risk of ischemic stroke in North Indian patients. Indian J EndocrinolMetab. 2014; 18: 721-725.
- Chaudhuri JR, Mridula KR, Alladi S, Anamika A, Umamahesh M, Balaraju B, et al. Serum 25-hydroxyvitamin d deficiency in ischemic stroke and subtypes in Indian patients. J Stroke. 2014; 16: 44-50.
- Bae JM, Kim EH. Human papillomavirus infection and risk of lung cancer in never-smokers and women: an ‘adaptive’ meta-analysis. Epidemiol Health. 2015; 37: e2015052.
- Bae JM, Kim EH. Dietary intakes of citrus fruit and risk of gastric cancer incidence: an adaptive meta-analysis of cohort studies. Epidemiol Health. 2016; 38: e2016034.
- Bae JM, Yoon BK. The role of menopausal hormone therapy in reducing all-cause mortality in postmenopausal women younger than 60 years: an adaptive meta-analysis. J Menoapausal Med. 2018; 24: 139-142.
- Bae JM. Narrative reviews. Epidemiol Health. 2014; 36: e2014018.
- Bae JM. A suggestion for quality assessment in systematic reviews of observational studies in nutritional epidemiology. Epidemiol Health. 2016; 38: e2016014.
- Kojima G, Bell C, Abbott RD, Launer L, Chen R, Motonaga H, et al. Low dietary vitamin D predicts 34-year incident stroke: the Honolulu Heart Program. Stroke. 2012; 43: 2163-2167.
- Michos ED, Reis JP, Post WS, Lutsey PL, Gottesman RF, Mosley TH, et al. 25-Hydroxyvitamin D deficiency is associated with fatal stroke among whites but not blacks: The NHANES-III linked mortality files. Nutrition. 2012; 28: 367- 371.
- Schierbeck LL, Rejnmark L, Tofteng CL, Stilgren L, Eiken P, Mosekilde L, et al. Vitamin D deficiency in postmenopausal, healthy women predicts increased cardiovascular events: a 16-year follow-up study. Eur J Endocrinol. 2012; 167: 553-560.
- Sun Q, Pan A, Hu FB, Manson JE, Rexrode KM. 25-Hydroxyvitamin D levels and the risk of stroke: a prospective study and meta-analysis. Stroke. 2012; 43: 1470-1477.
- Kühn T, Kaaks R, Teucher B, Hirche F, Dierkes J, Weikert C, et al. Plasma 25-hydroxyvitamin D and its genetic determinants in relation to incident myocardial infarction and stroke in the European prospective investigation into cancer and nutrition (EPIC)-Germany study. PLoS One. 2013; 8: e69080.
- Perna L, Schöttker B, Holleczek B, Brenner H. Serum 25-hydroxyvitamin D and incidence of fatal and nonfatal cardiovascular events: a prospective study with repeated measurements. J Clin Endocrinol Metab. 2013; 98: 4908-4915.
Download PDF