Dietary magnesium intake and risk of stroke: a meta-analysis of prospective studies 1 4

See corresponding editorial on page 269. Dietary magnesium intake and risk of stroke: a meta-analysis of prospective studies 1 4 Susanna C Larsson, Nicola Orsini, and Alicja Wolk ABSTRACT Background: Prospective studies of dietary magnesium intake in relation to risk of stroke have yielded inconsistent results. Objective: We conducted a dose-response meta-analysis to summarize the evidence regarding the association between magnesium intake and stroke risk. Design: Relevant studies were identified by searching PubMed and EMBASE from January 1966 through September 2011 and reviewing reference lists of retrieved articles. We included prospective studies that reported RRs with 95% CIs of stroke for 3 categories of magnesium intake. Results from individual studies were combined by using a random-effects model. Results: Seven prospective studies, with 6477 cases of stroke and 241,378 participants, were eligible for inclusion in the meta-analysis. We observed a modest but statistically significant inverse association between magnesium intake and risk of stroke. An intake increment of 100 mg Mg/d was associated with an 8% reduction in risk of total stroke (combined RR: 0.92; 95% CI: 0.88, 0.97), without heterogeneity among studies (P = 0.66, I 2 = 0%). Magnesium intake was inversely associated with risk of ischemic stroke (RR: 0.91; 95% CI: 0.87, 0.96) but not intracerebral hemorrhage (RR: 0.96; 95% CI: 0.84, 1.10) or subarachnoid hemorrhage (RR: 1.01; 95% CI: 0.90, 1.14). Conclusion: Dietary magnesium intake is inversely associated with risk of stroke, specifically ischemic stroke. Am J Clin Nutr 2012;95:362 6. INTRODUCTION Dietary magnesium intake has been inversely associated with risk factors for stroke such as hypertension (1, 2), metabolic syndrome (3), and type 2 diabetes (2, 4). Randomized clinical trials have shown that magnesium supplementation modestly reduces diastolic blood pressure (5) as well as fasting C-peptide and insulin concentrations (6). Furthermore, findings from animal studies have shown that high-magnesium diets have favorable effects on plasma glucose and blood lipid concentrations (7), and magnesium deficiency increases the susceptibility of lipoproteins to peroxidation (8). Hence, a high dietary magnesium intake may potentially reduce risk of stroke. To our knowledge, the epidemiologic evidence on the relation between dietary magnesium intake and risk of stroke has not yet been summarized. Therefore, we performed a systematic review and dose-response meta-analysis to assess the association between magnesium intake and risk of total stroke and stroke subtypes. METHODS Literature search and selection We followed standard criteria for conducting and reporting of meta-analyses of observational studies (9). We conducted a literature search of PubMed (http://www.ncbi.nlm.nih.gov/ pubmed) and EMBASE (http://www.embase.com) from January 1966 through September 2011 by using the key words magnesium intake combined with stroke. In addition, we reviewed reference lists of retrieved articles to identify additional relevant studies. No language restrictions were imposed. Studies were included in this meta-analysis if they fulfilled the following criteria: 1) had a prospective design, 2) the exposure of interest was magnesium intake, 3) the outcome was stroke, and 4) the studies reported RRs with 95% CIs for 3 quantitative categories of magnesium intake. RRs had to at least be adjusted for age and sex (if applicable). Data extraction The following data were extracted from each study: first author s last name, publication year, name of the cohort study, study location, years of follow-up, sex, age, sample size (number of cases and total number of participants), magnesium intake categories, covariates adjusted for in the multivariable analysis, and RRs with their 95% CIs for each category of magnesium intake. We extracted RRs that reflected the greatest degree of adjustment for potential confounders. Data extraction was conducted independently by 2 investigators (SCL and NO) with disagreements resolved by consensus. 1 From the Division of Nutritional Epidemiology, National Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden. 2 Funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the manuscript. 3 Supported by a research grant from the Swedish Council for Working Life and Social Research and a Research Fellow grant from Karolinska Institutet (SCL). 4 Address correspondence to SC Larsson, Division of Nutritional Epidemiology, National Institute of Environmental Medicine, Karolinska Institutet, Box 210, SE-17177 Stockholm, Sweden. E-mail: susanna.larsson@ki.se. Received June 23, 2011. Accepted for publication October 24, 2011. First published online December 28, 2011; doi: 10.3945/ajcn.111.022376. 362 Am J Clin Nutr 2012;95:362 6. Printed in USA. Ó 2012 American Society for Nutrition

MAGNESIUM INTAKE AND STROKE 363 Statistical analysis Because the range of magnesium intake and the cutoffs for the categories differed between studies, we estimated an RR with 95% CI for an increase in intake of 100 mg Mg/d for each study. We used the method proposed by Greenland and Longnecker (10) and Orsini et al (11) to compute the trend from the correlated log RR estimates across categories of magnesium intake. For each study, the median or mean magnesium intake for each category was assigned to each corresponding RR. When the median or mean intake per category was not provided, we assigned the midpoint of upper and lower boundaries in each category as the average intake. If the lower or upper boundary for the lowest and highest category, respectively, was not reported, we assumed that the boundary had the same amplitude as the closest category. To examine a potential nonlinear association between magnesium intake and stroke risk, we performed a 2-stage, random-effects, dose-response meta-analysis. This was done by modeling magnesium intake by using restricted cubic splines with 3 knots at fixed percentiles 10%, 50%, and 90% of the distribution (12). First, a restricted cubic-spline model was estimated by using generalized least-square regression and taking into account the correlation within each set of published RRs as described by Orsini et al (11). Second, we combined the study-specific estimates by using the restricted maximum likelihood method in a multivariate random-effects metaanalysis (13). A P value for nonlinearity was calculated by testing the null hypothesis that the coefficient of the second spline was equal to zero. We showed no significant departure from a linear-response model (P-spline transformation = 0.15). Statistical heterogeneity among studies was evaluated by using I 2 statistics (14). Three cutoffs of these I 2 values were considered as follows:,30% (no or marginal between-study heterogeneity), 30 75% (mild heterogeneity), and.75% (notable heterogeneity). We performed analyses stratified by study location and stroke subtypes. Publication bias was assessed with Egger s regression test (15). All statistical analyses were conducted with Stata software (version 10.1; StataCorp). P, 0.05 was considered statistically significant. RESULTS Study characteristics Our literature search identified 7 independent prospective studies of dietary magnesium intake and stroke risk (Figure 1) (2, 3, 16 20). The studies were published between 1998 and 2011 and included a total of 6477 stroke cases and 241,378 participants (Table 1). Four studies were conducted in the United States, 2 studies were conducted in Europe, and one study was conducted in Taiwan. The median magnesium intake in the study populations ranged from 242 mg/d (in US men and women) to 471 mg/d (in Finnish men). All studies provided risk estimates that were adjusted for age, smoking, and BMI. Most studies also controlled for physical activity (6 studies), diabetes (6 studies), history of hypertension or measured blood pressure (6 studies), and alcohol consumption (6 studies); fewer studies adjusted for other nutrients (3 studies). FIGURE 1. Flowchart of study selection by using PubMed (http://www. ncbi.nlm.nih.gov/pubmed) and EMBASE (http://www.embase.com). Magnesium intake and stroke Estimated RRs of total stroke for an increase in intake of 100 mg Mg/d for each of the included studies and the combined estimate are shown in Figure 2. The combined RR of total stroke was 0.92 (95% CI: 0.88, 0.97) without heterogeneity among studies (P = 0.66, I 2 = 0%). In a sensitivity analysis in which one study at the time was omitted and the rest of studies were analyzed, combined RRs ranged from 0.91 (95% CI: 0.85, 0.97) to 0.93 (95% CI: 0.88, 0.99). Hypertension and diabetes were potential confounders of the association between magnesium intake and stroke risk. The results persisted when we excluded the study (17) that did not adjust for both hypertension and diabetes (RR: 0.92; 95% CI: 0.87, 0.98). We showed no evidence of publication bias (Egger s test: P = 0.54). Combined RRs of stroke for an increment in intake of 100 mg Mg/d were 0.91 (95% CI: 0.85, 0.96) in studies conducted in the United States and 0.95 (95% CI: 0.88, 1.02) in the European studies. Four studies provided results for stroke subtypes (3, 17, 18, 20) and 2 studies reported results for ischemic stroke only (2, 19). For these 6 studies, combined RRs for an increase in intake of 100 mg Mg/d were 0.91 (95% CI: 0.87, 0.96) for ischemic stroke, 0.96 (95% CI: 0.84, 1.10) for intracerebral hemorrhage, and 1.01 (95% CI: 0.90, 1.14) for subarachnoid hemorrhage. DISCUSSION This meta-analysis of 7 prospective studies showed a significant inverse association between dietary magnesium intake and risk of stroke. An increase in intake of 100 mg Mg/d was associated with an 8% and 9% decreased risk of total stroke and ischemic stroke, respectively. Magnesium intake was not associated with hemorrhagic strokes. However, the number of cases was lower for hemorrhagic stroke than for ischemic and total strokes, which led to a lower statistical power to estimate the association between magnesium intake and risk of hemorrhagic stroke. Rich food sources of magnesium are green leafy vegetables, nuts, beans, and whole grains. The US Recommended Dietary Allowance of magnesium for women and men 31 70 y of age is

364 LARSSON ET AL TABLE 1 Prospective studies of magnesium intake and risk of stroke 1 Study, cohort study name (country) No. of cases (cohort size) Years of followup Sex, age Magnesium intake [RR (95% CI) for highest compared with lowest category of intake] Adjustments Ascherio et al, 1998 (16), Health Professionals Follow-up Study (United States) Iso et al, 1999 (17), Nurses Health Study Song et al, 2005 (3), Women s Health Study Larsson et al, 2008 (18), Alpha-Tocopherol, Beta- Carotene Cancer Prevention Study (Finland) Weng et al, 2008 (19), CardioVascular Disease risk FACtor Two-township Study (Taiwan) Ohira et al, 2009 (2), Atherosclerosis Risk in Communities Study Larsson et al, 2011 (20), Swedish Mammography Cohort (Sweden) 328 total strokes (43,738 nondiabetics) 690 total strokes, 386 ISs, 74 ICHs, and 129 SHs (85,764) 8 M, 40 75 y Median: 452 compared with 243 mg/d [0.92 (0.58, 1.46)] 14 F, 34 59 y Median: 381 compared with 211 mg/d [0.80 (0.63, 1.01)] 368 total strokes (35,601) 10 F, 39 89 y Median: 433 compared with 255 mg/d [0.90 (0.65, 1.26)] 2702 ISs, 383 ICHs, and 196 SHs (26,556) 132 ISs (1772) 10.6 M and F, 40 y 577 ISs (13,277) 15 M and F, 45 64 y 1680 total strokes, 1310 ISs, 154 ICHs, and 79 SHs (34,670) 13.6 M, 50 69 y Median: 575 compared with 382 mg/d [IS: 0.85 (0.76, 0.97)].282.2 compared with,242.6 mg/d [0.68 ( 0.45, 1.04)].307 compared with,186 mg/d [0.80 (0.75, 1.13)] 10.4 F, 49 83 y Median: 373 compared with 267 mg/d [1.02 (0.82, 1.27)] Age, smoking, profession, histories of hypertension and hypercholesterolemia, family history of MI, BMI, physical activity, and intakes of alcohol, dietary fiber, potassium, and total energy Age, smoking, menopausal status, postmenopausal hormone use, BMI, exercise, histories of diabetes and high cholesterol, aspirin use, multivitamin use, vitamin E use, and intakes of alcohol, omega-3 fatty acids, and calcium Age, randomized treatment assignment, smoking, BMI, exercise, postmenopausal hormone use, multivitamin use, histories of diabetes, hypertension, and hypercholesterolemia, family history of MI, intake of alcohol, and total energy Age, supplementation group, cigarettes smoked daily, BMI, physical activity, systolic and diastolic BP, serum total and HDL cholesterol, histories of diabetes and ischemic heart disease, intake of alcohol, and total energy Age, sex, age sex, smoking, sex smoking, area, central obesity, BMI, diabetes, physical activity, hypertension, use of antihypertensive drugs, selfreported heart disease, hypercholesterolemia, hypertriglyceridemia, fibrinogen, apolipoprotein B, plasminogen, and alcohol intake Age, sex, center, education, smoking, BMI, diabetes, systolic blood pressure, use of antihypertensive medication, LDL and HDL cholesterol, fibrinogen, von Willebrand factor, and total energy intake Age, education, smoking, BMI, physical activity, history of diabetes, history of hypertension, aspirin use, family history of myocardial infarction, and intakes of alcohol, protein, cholesterol, total fiber, folate, and total energy 1 BP, blood pressure; ICH, intracerebral hemorrhage; IS, ischemic stroke; MI, myocardial infarction; SH, subarachnoid hemorrhage.

MAGNESIUM INTAKE AND STROKE 365 FIGURE 2. RRs of total stroke for an increment in intake of 100 mg Mg/ d. Squares represent study-specific RR estimates [the size of the square reflects the study-specific statistical weight (ie, the inverse of the variance)], horizontal lines represent 95% CIs, and the diamond represents the combined RR estimate with its 95% CI. 320 and 420 mg/d, respectively (21). The approximate magnesium content of some foods is 157 mg in 1 cup cooked spinach, 77 mg in 1 oz roasted cashew nuts, 134 mg in 1 cup canned white beans, 88 mg in 1 cup cooked oat bran, and 84 mg in 1 cup cooked brown rice (22). There are several potential mechanisms by which magnesium may reduce risk of ischemic stroke. The protective effect of magnesium against risk of stroke may in part be related to its blood pressure lowering effect. Findings from a meta-analysis of 12 randomized clinical trials (median follow-up: 11 wk) showed that magnesium supplementation (mean dose: ;400 mg/d) may slightly decrease diastolic blood pressure by 2.2 mm Hg (5). Most studies included in this meta-analysis controlled for history of hypertension or blood pressure levels. There may have been overadjustment if the association between magnesium intake and stroke was mediated via a reduction in blood pressure. Magnesium intake has also been inversely associated with risk of type 2 diabetes (2, 4), which is a risk factor for ischemic stroke (23, 24). Therefore, it is possible that type 2 diabetes is a mediator of the association between magnesium intake and ischemic stroke. Moreover, a high-magnesium diet has been shown to have favorable effects on plasma glucose and blood lipid concentrations in rats with chronic diabetes (7). In addition, there are reports that showed that magnesium deficiency increased the susceptibility of lipoproteins to peroxidation in animals (8). This meta-analysis had several strengths. First, our quantitative assessment was based on data from prospective studies, which eliminated recall and selection biases. Furthermore, by combining results from several studies, we had high statistical power to estimate the relation between magnesium intake and risk of stroke. Also, all studies were very consistent with no heterogeneity among study results. This meta-analysis also had several potential limitations. First, as a meta-analysis of observational studies, the possibility of confounding as a potential explanation for the observed inverse association between magnesium intake and stroke risk could not be excluded. Although the majority of studies adjusted for known risk factors for stroke, such as age, cigarette smoking, BMI, physical activity, histories of diabetes and hypertension, and alcohol consumption, the possibility for residual confounding remained. Moreover, it could not be excluded that other nutrients or dietary components that are correlated with dietary magnesium may have been responsible, partly or in whole, for the observed association. A second limitation was the misclassification of magnesium intake, which would most likely have led to an underestimation of the true relation between magnesium intake and stroke. Finally, in a metaanalysis of published studies, publication bias may be a problem. However, we showed no evidence of publication bias in this metaanalysis. In conclusion, results from this meta-analysis indicate that dietary magnesium intake is inversely associated with risk of stroke. Although it is premature to recommend magnesium supplementation to reduce risk of stroke, increased consumption of magnesium-rich foods such as green leafy vegetables, beans, nuts, and whole-grain cereals appears to be prudent. The authors responsibilities were as follows SCL: provided study oversight and wrote and took primary responsibility for the final content of the manuscript; SCL and NO: undertook data collection and performed statistical analyses; and all authors: designed the research, assisted in the interpretation of analyses and revision of the manuscript, and read and approved the final manuscript. None of the authors had a conflict of interest. REFERENCES 1. Ma J, Folsom AR, Melnick SL, Eckfeldt JH, Sharrett AR, Nabulsi AA, Hutchinson RG, Metcalf PA. Associations of serum and dietary magnesium with cardiovascular disease, hypertension, diabetes, insulin, and carotid arterial wall thickness: the ARIC study. Atherosclerosis Risk in Communities Study. J Clin Epidemiol 1995;48:927 40. 2. Ohira T, Peacock JM, Iso H, Chambless LE, Rosamond WD, Folsom AR. Serum and dietary magnesium and risk of ischemic stroke: the Atherosclerosis Risk in Communities Study. Am J Epidemiol 2009; 169:1437 44. 3. Song Y, Manson JE, Cook NR, Albert CM, Buring JE, Liu S. Dietary magnesium intake and risk of cardiovascular disease among women. Am J Cardiol 2005;96:1135 41. 4. Larsson SC, Wolk A. Magnesium intake and risk of type 2 diabetes: a meta-analysis. J Intern Med 2007;262:208 14. 5. Dickinson HO, Nicolson DJ, Campbell F, Cook JV, Beyer FR, Ford GA, Mason J. Magnesium supplementation for the management of essential hypertension in adults. Cochrane Database Syst Rev 2006;3: CD004640. 6. Chacko SA, Sul J, Song Y, Li X, LeBlanc J, You Y, Butch A, Liu S. Magnesium supplementation, metabolic and inflammatory markers, and global genomic and proteomic profiling: a randomized, doubleblind, controlled, crossover trial in overweight individuals. Am J Clin Nutr 2011;93:463 73. 7. Soltani N, Keshavarz M, Dehpour AR. Effect of oral magnesium sulfate administration on blood pressure and lipid profile in streptozocin diabetic rat. Eur J Pharmacol 2007;560:201 5. 8. Rayssiguier Y, Gueux E, Bussiere L, Durlach J, Mazur A. Dietary magnesium affects susceptibility of lipoproteins and tissues to peroxidation in rats. J Am Coll Nutr 1993;12:133 7. 9. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, Moher D, Becker BJ, Sipe TA, Thacker SB. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Metaanalysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000;283:2008 12. 10. Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol 1992;135:1301 9. 11. Orsini N, Bellocco R, Greenland S. Generalized least squares for trend estimation of summarized dose-response data. Stata Journal 2006;6: 40 57. 12. Harrell FE Jr, Lee KL, Pollock BG. Regression models in clinical studies: determining relationships between predictors and response. J Natl Cancer Inst 1988;80:1198 202.

366 LARSSON ET AL 13. Jackson D, White IR, Thompson SG. Extending DerSimonian and Laird s methodology to perform multivariate random effects metaanalyses. Stat Med 2010;29:1282 97. 14. Higgins JP, Thompson SG. Quantifying heterogeneity in a metaanalysis. Stat Med 2002;21:1539 58. 15. Egger M, Davey Smith G, Schneider M, Minder C. Bias in metaanalysis detected by a simple, graphical test. BMJ 1997;315:629 34. 16. Ascherio A, Rimm EB, Hernan MA, Giovannucci EL, Kawachi I, Stampfer MJ, Willett WC. Intake of potassium, magnesium, calcium, and fiber and risk of stroke among US men. Circulation 1998;98:1198 204. 17. Iso H, Stampfer MJ, Manson JE, Rexrode K, Hennekens CH, Colditz GA, Speizer FE, Willett WC. Prospective study of calcium, potassium, and magnesium intake and risk of stroke in women. Stroke 1999;30: 1772 9. 18. Larsson SC, Virtanen MJ, Mars M, Männistö S, Pietinen P, Albanes D, Virtamo J. Magnesium, calcium, potassium, and sodium intakes and risk of stroke in male smokers. Arch Intern Med 2008;168:459 65. 19. Weng LC, Yeh WT, Bai CH, Chen HJ, Chuang SY, Chang HY, Lin BF, Chen KJ, Pan WH. Is ischemic stroke risk related to folate status or other nutrients correlated with folate intake? Stroke 2008;39:3152 8. 20. Larsson SC, Virtamo J, Wolk A. Potassium, calcium, and magnesium intake and risk of stroke in women. Am J Epidemiol 2011;174:35 43. 21. Dietary Reference Intakes for calcium, phosphorus, magnesium, vitamin D, and fluoride. Washington, DC: National Academy of Sciences. Institute of Medicine. Food and Nutrition Board, 1997. 22. US Department of Agriculture, Agricultural Research Service. USDA national nutrient database for standard reference, release 24, 2011. Available from: http://www.ars.usda.gov/ba/bhnrc/ndl. 23. Andersen KK, Olsen TS, Dehlendorff C, Kammersgaard LP. Hemorrhagic and ischemic strokes compared: stroke severity, mortality, and risk factors. Stroke 2009;40:2068 72. 24. Janghorbani M, Hu FB, Willett WC, Li TY, Manson JE, Logroscino G, Rexrode KM. Prospective study of type 1 and type 2 diabetes and risk of stroke subtypes: the Nurses Health Study. Diabetes Care 2007;30: 1730 5.