Abstract
Objective:
To investigate the longitudinal association between magnesium (Mg) intake and the risk of metabolic syndrome (MetS).
Methods:
Poisson regression models with robust standard error estimation were used to examine the association between total Mg intake and the risk of MetS in 6,802 participants aged ≥45 years at baseline in the REasons for Geographic And Racial Differences in Stroke (REGARDS) study. Dietary data were collected using the modified Block 98 food frequency questionnaire (FFQ) at baseline and incident MetS was diagnosed during follow-up if a participant had three or more of the five components of MetS based on the harmonized definition.
Results:
A total of 1,470 participants developed MetS during an average follow-up of 10 years. Comparing the highest quintile of total Mg intake (>437.9 mg/day) to the lowest group (<223.5 mg/day), total Mg intake had a significant inverse association with the risk of MetS [relative risk (RR) = 0.79 (0.63, 0.98), P trend = 0.043]. Dietary Mg intake was inversely associated with MetS [RR = 0.72 (0.56, 0.91), P trend = 0.006]. Adjusting for baseline components of MetS attenuated the associations, but the linear trends remained.
Conclusion:
The findings from this study indicate that dietary Mg intake was inversely associated with the risk of MetS. We recommend further studies to explain the underlying mechanisms of action.
Keywords: Magnesium intake, metabolic syndrome, cohort study, diet, REGARDS
Introduction
Metabolic syndrome (MetS) is a constellation of conditions that are risk factors for cardiovascular diseases and other chronic diseases 1; these risk factors include high blood pressure, abnormal lipid profile, central obesity, and high fasting blood glucose levels. About one-third of the general population in the United States has MetS in 2011 to 2012 according to the National Health and Nutritional Examination Survey (NHANES) 2.
Several studies have linked magnesium (Mg) deficiency with the components of MetS. Evidence from meta-analyses of clinical trials indicated Mg supplementation modestly reduced blood pressure 3–5. Dietary Mg intake is also inversely associated with the risk of insulin resistance 6–8, inversely associated with central obesity, directly associated with high-density lipoprotein (HDL) but not associated with triglyceride 9. Recent meta-analyses 10,11 of mostly cross-sectional studies and two cohort studies 12,13 suggest an inverse association between dietary Mg intake and the risk of MetS. Earlier, a review by He et al suggested Mg intake and serum Mg are inversely associated with MetS 14.
Longitudinal studies on the association between Mg intake and the risk of MetS are scant. This study aimed to investigate the long-term association of Mg intake with MetS. A second purpose was to determine if the associations were different by sex, race, and regions using the data from the REGARDS study. Our research hypothesis was that Mg intake is inversely associated with the risk of MetS.
Methods
Study participants
The REGARDS study is a longitudinal population-based study designed to investigate risk factors associated with excess stroke mortality among black participants and residents of Stroke Belt regions that include North Carolina, South Carolina, Georgia, Tennessee, Mississippi, Alabama, Louisiana, and Arkansas 15. The REGARDS study sample is randomly selected from commercially available nationwide lists purchased from GeneSys, Inc. The participants were aged ≥45 years at baseline, and self-identified as black or white. The participants (n=30,239) were enrolled between January 2003 and October 2007, Supplemental Figure S1 A sample balanced by race (42% black) and sex (55% female) was recruited (21% of all participants) from the Stroke Buckle (southern regions with the highest stroke mortality 16), from the general Stroke Belt (35% of all participants) and from the remaining 40 contiguous states (44% of all participants).
The REGARDS cohort was described in detail by Howard et al 15,17. Briefly, potential participants were mailed letters introducing the study and were contacted by phone 2 weeks later. Verbal consent of participants was obtained; then, demographic, and behavioral data were collected using computer-assisted telephone interview. Examination Management Services, Inc. (Irving, Texas, USA), trained technicians obtained written informed consent, physical measurements, and a medication inventory during a subsequent in-home visit from participants that fasted 10 to 12 hours overnight. The rest of the data were collected by self-administered questionnaires. The second in-home visit was conducted in late 2016 with the same protocols.
Potential participants were excluded if they were of races other than black and non-Hispanic white, undergoing active treatment for cancer, had medical conditions that prevented long-term participation, displayed cognitive impairment as judged by the trained interviewers, resided in or were included on a waiting list for nursing homes, or were not able to communicate in English. For the current study, we further excluded records that had missing data for Calcium (Ca) intake, Mg intake, components of MetS including blood pressure (BP), fasting blood glucose, HDL, blood triglycerides (TG), or waist circumference (WC). No participant had implausible energy intake (< 500 kcal/day or > 5000 kcal/day) 18. Of the 15,972 participants who completed the first and the second in-home visits, 6,581 participants had baseline MetS, 2,589 participants had missing values for Mg (n = 1,651), or other MetS definition variables (n = 938) leaving 6,802 participants in the present analyses, Figure 1. There was no statistically significant difference in mean age between participants that had missing values and those included in the analysis, P = 0.079. The participants excluded from the analysis due to missing values had 2.23 mg/dL higher baseline LDL, P = 0.003, but a borderline difference in follow-up LDL, P = 0.055, no difference in baseline triglyceride (TG), P = 0.751, but those excluded had a 3.44 mg/dL lower follow-up TG, P = 0.027 and lower baseline systolic blood pressure (SBP) and diastolic blood pressure (DBP), P < 0.0001. Those excluded had higher baseline total calorie intake P = 0.040 and blood glucose, P = 0.007, and higher follow-up glucose, P = 0.045. The IRBs of Indiana University at Bloomington approved the current study.
Figure 1.
Study participants selection process
Exposure Measurement
Dietary data were collected using the self-administered modified Block 98 FFQ with over 150 multiple choice questions on 107 food items18–20. Nutrients in foods were extracted by NutritionQuest using proprietary algorithms based on the USDA Database 21. The amount of each food consumed was calculated by multiplying the reported frequency by the portion size for each food item. The amount of a nutrient from each food was derived by multiplying the amount consumed by the amount of the nutrient in the given FFQ line item. Nutrients were summed over all FFQ food items to provide estimates of total daily nutrient intake. The Block FFQ has been validated for the assessment of nutrients, including Mg and calcium 22,23. Total Mg included dietary Mg and supplemental Mg. For total Mg intake, the deattenuated correlation of FFQ with an average of two 24-hour recalls was 0.63 ( 95% CI: 0.34, 0.92) among Canadian women and for total calcium the correlation was 0.71 (0.35, 1.00) 22. In the other validation study, Block FFQ for Mg intake had an energy adjusted deattenuated correlation of 0.81 with repeat 24-hour dietary recalls over a year among women and 0.76 among men. Calcium had a correlation of 0.66 among women and 0.72 among men23. Supplemental Mg intake was assessed using a medication inventory whereby the participants showed trained health professionals any medication they used at least once in the 2 weeks prior to the first in-home visit, and then supplemental Mg was estimated from actual Mg supplements, multivitamins or Mg-containing medications.
Outcome Ascertainment
Participants in the REGARDS Study were diagnosed with incident MetS if they had three or more of the five components of MetS during the follow-up, including SBP ≥ 130 mmHg, and/or DBP ≥ 85 mmHg or treatment with anti-hypertensive medication; hyperglycemia (fasting blood glucose level ≥ 100 mg/dL) or treatment with glucose-lowering medication; TG (≥ 1.7 mmol/L) or treatment for elevated TG; HDL (< 1.3 mmol/L or < 50 mg/dL in women and < 1.0 mmol/L or < 40 mg/dL in men) or treatment for reduced HDL; and central obesity (WC ≥ 88 cm in women and ≥ 102 cm in men) according to the harmonized definition of MetS 24. The harmonized definition incorporates the National Cholesterol Education Program Adult Treatment Panel III (ATP III) MetS definition but lowered the threshold for fasting blood glucose to 100 mg/dL.
Two measurements of BP were taken by aneroid sphygmomanometer during each home visit and the average values were recorded. In REGARDS study, WC was measured after a participant exhaled, over the skin or lightweight clothing at the midpoint between the lowest rib and the top of the iliac crest using a cloth tape measure. REGARDS data for blood glucose, TG, and cholesterol were collected after 10 to 12 hours fasting overnight by venipuncture samples during home visits.
Covariates
The REGARDS data included demographics such as age (continuous), sex, race (blacks or non-Hispanic whites), regions (Stroke Belt, Stroke Buckle or Non-Belt regions) and socioeconomic status (education and income-categorical), aspirin use (yes/no), cigarette smoking (never, past or current), alcohol intake (heavy (greater than 1 drink per day for women and 2 drinks per day for men), moderate (1 drink per day for women or 2 drinks per day for men) or none), calcium intake (continuous), and physical activity level (≥ 4 times/week, 1–3 times/week or none).
Statistical Analysis
Analysis of variance was used for normally distributed continuous variables and Kruskal-Wallis test was used for non-normally distributed continuous variables or for ordinal variables to compare across the quintiles of Mg intake. Chi-square test was used to examine the association of categorical variables with nutrient quintiles. Given that the exact time of MetS was unknown in most cases and there was a wide interval censoring, survival analysis was not conducted. Thus, Poisson regression models with robust standard error estimation method were used to estimate the relative risk (RR) of incident MetS by comparing the higher quintiles to the lowest quintile of Mg intake in a sequential manner. Model 1 was adjusted for age, sex, race, and region. Model 2 included variables in model 1 and further adjusted for baseline and income, education, physical activity, smoking status, and alcohol intake. Model 3 included variables in model 2, regular aspirin use (baseline), total energy intake, Ca intake, and low-density lipoprotein (baseline and 2nd home-visit). The interaction between age and race was also adjusted whenever significant in the models. In a sensitivity analysis, we further adjusted for baseline for continuous forms of the individual components of MetS. The possible non-linear association of Mg intake with MetS was evaluated by restricted cubic splines analysis with 4 knots using SAS macro “lgtphcurv9” 25. The median of the first quintile was used as the reference for the restricted cubic spline analysis. Further, to assess whether there was a difference in the associations of interest by sex, race or regions, stratified analyses were conducted by each of these pre-specified variables. The associations between dietary and total Mg intake and the risk of the individual components of MetS (binary according to the harmonized definition of MetS) were also conducted adjusting for the covariates in Model 3 for MetS. To examine the relations between each component of MetS (in continuous form) and the MetS, we created a MetS score and conducted a correlation and path analysis.
Results
A total of 6,802 REGARDS participants free of MetS at baseline were followed for a duration ranging from 9 to 13 years, and 1,470 participants met the criteria of MetS during the follow-up, Table 1. The participants with the highest Mg intake tended to be older, male, white, more likely to be engaged in physical activity, and more to be categorized as moderate drinkers. The participants with the highest Mg intake were also more likely to be non-smokers; they were more likely to have lower BMI, and DBP. Those in the highest quintile of total Mg intake had about twice total calorie intake as those in the first quintile but also had a better BMI. This could be explained by the higher physical activity among participants with the highest total Mg intake.
Table 1 :
Baseline characteristics of study participants by total Mg quintiles (n=6,802)
| Total Mg intake quintiles | |||||||
|---|---|---|---|---|---|---|---|
| Quintile 1 | Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 | P | ||
| Total Mg (mg/day) | 169.3 (38.0) | 259.1 (19.6) | 324.1(18.8) | 394.4 (23.2) | 541.3 (97.7) | ||
| Dietary Mg (mg/day) | 160.0 (39.5) | 225.5 (45.5) | 266.8 (48.9) | 327.1 (47.7) | 469.2 (104.5) | <0.0001 | |
| Mg supplement (mg/day) | 9.3 (25.0) | 36.7 (44.3) | 57.3 (46.2) | 67.3 (43.5) | 72.1 (41.5) | <0.0001 | |
| Total Ca (mg/day) | 617.1 (420.8) | 895.4 (459.4) | 1,087.0 (508.3) | 1,254.4 (531.5) | 1,550.9 (617.5) | <0.0001 | |
| Dietary Ca (mg/day) | 391.0 (182.9) | 541.7 (223.7) | 643.8 (252.1) | 776.6 (284.0) | 1056.1 (396.2) | <0.0001 | |
| Age (years) | 62.6 (9.1) | 62.5 (8.7) | 63.7 (8.4) | 63.4 (8.6) | 63.4 (8.7) | <0.0001 | |
| Sex (% female) | 63.3 | 59.6 | 58.0 | 53.2 | 48.3 | <0.0001 | |
| Race (% black) | 40.2 | 28.8 | 22.0 | 19.6 | 17.2 | <0.0001 | |
| Education | Less than high school | 8.9 | 4.6 | 3.3 | 4.0 | 2.8 | <0.0001 |
| High school graduate | 26.0 | 21.2 | 20.4 | 18.4 | 15.8 | ||
| Some college | 28.3 | 28.6 | 25.3 | 23.5 | 22.6 | ||
| College graduate and above | 37.1 | 45.6 | 51.0 | 54.0 | 58.8 | ||
| BMI (mean, SD) | 27.2 (5.1) | 26.8 (4.8) | 26.6 (4.9) | 26.5 (4.7) | 26.2 (4.4) | <0.0001 | |
| Physical Activity | None | 32.7 | 26.0 | 24.9 | 24.0 | 18.2 | <0.0001 |
| 1-3times /week | 38.7 | 43.5 | 40.5 | 39.4 | 39.2 | ||
| ≥4 times/week | 28.6 | 30.5 | 34.7 | 36.6 | 42.7 | ||
| Alcohol | Heavy | 4.5 | 6.0 | 5.5 | 5.9 | 5.4 | <0.0001 |
| Moderate | 35.7 | 42.2 | 43.5 | 47.7 | 47.9 | ||
| None | 59.7 | 51.7 | 46.5 | 46.5 | 46.7 | ||
| Smoking | Current | 14.1 | 12.6 | 9.6 | 8.9 | 6.9 | <0.0001 |
| Past | 33.8 | 36.4 | 38.9 | 40.3 | 42.1 | ||
| Never | 52.1 | 51.0 | 51.6 | 50.8 | 51.0 | ||
| Aspirin use (%) | 29.4 | 35.0 | 40.6 | 42.0 | 43.0 | <0.0001 | |
| Calorie intake (kcal/day) | 1,139.4 (357.1) | 1,445.7 (437.0) | 1,645.7 (467.7) | 1,912.8 (544.5) | 2,458.5 (696.9) | <0.0001 | |
| FBG (mg/dL) | 91.7 (18.4) | 91.3 (16.3) | 91.1 (15.0) | 91.0 (16.1) | 90.4 (13.1) | 0.147 | |
| HDL (mg/dL) | 58.4 (15.8) | 58.4 (16.6) | 58.4 (16.2) | 58.1 (16.3) | 57.2 (15.9) | 0.072 | |
| LDL (mg/dL) | 119.7 (33.1) | 117.2 (32.1) | 114.3 (33.2) | 116.8 (32.1) | 114.1 (30.7) | <0.0001 | |
| TG (mg/dL) | 105.3 (51.0) | 108.4 (54.8) | 112.7 (58.0) | 111.4 (32.1) | 104.2 (83.8) | <0.0001 | |
| Total cholesterol (mg/dL) | 199.4 (37.1) | 197.4 (36.6) | 195.5 (37.5) | 197.2 (35.9) | 191.9 (34.7) | <0.0001 | |
| SBP (mmHg) | 122.8 (15.8) | 121.9 (14.7) | 121.8 (15.4) | 121.4 (14.5) | 121.2 (14.4) | 0.289 | |
| DBP (mmHg) | 75.1 (9.4) | 75.5 (8.8) | 74.7 (8.7) | 74.3 (8.6) | 74.2 (8.7) | <0.0001 | |
| WC (cm) | 88.1 (12.8) | 88.2 (14.7) | 88.1 (13.2) | 88.1 (12.5) | 88.3 (12.9) | 0.994 | |
Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; FBG, fasting blood glucose; HDL, high density lipoprotein; SBP, systolic blood pressure; SD, standard deviation; TG, triglycerides; WC, waist circumference
The values are means (SD) for continuous variables and percent (%) for categorical variables. Chi-squared test for categorical variables, and ANOVA and Kruskal-Wallis test for continuous variables were used to compute the P values.
When the highest quintile of total Mg was compared to the lowest, total Mg intake had a borderline non-significant inverse association with MetS (RR = 0.79 (0.63, 0.98), Ptrend = 0.043), Table 2. Dietary Mg intake was inversely associated with MetS (RR = 0.72 (0.56, 0.91), P trend = 0.006) comparing the highest quintile to the lowest, Table 2. In sensitivity analysis, adjusting for baseline components of MetS attenuated the associations of total Mg with MetS, but the linear trend remained significant, Table 2. In stratified analyses by sex, race, and region, the association for both total and dietary Mg remained among female participants, and for dietary Mg among white participants and participants from the Stroke Belt region (data not shown) but disappeared the other strata, Supplemental Table S3. Restricted cubic spline analyses indicated no non-linear pattern of association, the P for non-linear association were 0.388 for total Mg and 0.296 for dietary Mg, Supplemental Figure 2 & 3.
Table 2:
The RR (95% CI) of MetS by Mg quintiles (n=6,802)
| Quintile 1 | Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 | a Ptrend | |
|---|---|---|---|---|---|---|
| Total Mg, mg/day | <223.50 | 223.50–291.20 | 291.30–357.10 | 357.20–437.90 | >437.90 | |
| Number of cases | 308 | 291 | 304 | 284 | 283 | |
| Number of participants | 1360 | 1361 | 1361 | 1362 | 1358 | |
| Model 1 | 1.00 | 0.96 (0.83, 1.11) | 1.02 (0.89, 1.17) | 0.95 (0.83, 1.10) | 0.95 (0.83, 1.10) | 0.541 |
| Model 2 | 1.00 | 0.98 (0.85, 1.13) | 1.04 (0.90, 1.20) | 0.98 (0.85, 1.14) | 1.00 (0.86, 1.16) | 0.699 |
| Model 3 | 1.00 | 0.88 (0.75, 1.03) | 0.90 (0.77, 1.07) | 0.83 (0.69, 1.00) | 0.79 (0.63, 0.98) | 0.043 |
| Sensitivity Analysis | 1.00 | 0.89 (0.76, 1.04) | 0.91 (0.77, 1.07) | 0.85 (0.71, 1.02) | 0.80 (0.64, 1.00) | 0.064 |
| Dietary Mg, mg/day | <187.50 | 187.5–243.00 | 243.10–298.40 | 298.50–375.10 | >375.10 | |
| Number of cases | 301 | 299 | 310 | 275 | 285 | |
| Number of participants | 1,361 | 1,360 | 1,360 | 1,361 | 1,360 | |
| Model 1 | 1.00 | 1.02 (0.88, 1.17) | 1.06 (0.92, 1.22) | 0.95 (0.82, 1.10) | 0.98 (0.85, 1.13) | 0.513 |
| Model 2 | 1.00 | 1.00 (0.87, 1.16) | 1.06 (0.92, 1.22) | 0.96 (0.83, 1.12) | 1.00 (0.86, 1.16) | 0.823 |
| Model 3 | 1.00 | 0.90 (0.77, 1.05) | 0.89 (0.75, 1.05) | 0.76 (0.63, 0.92) | 0.72 (0.56, 0.91) | 0.006 |
| Sensitivity Analysis | 1.00 | 0.92 (0.79, 1.08) | 0.91 (0.77, 1.07) | 0.80 (0.66, 0.96) | 0.76 (0.60, 0.97) | 0.018 |
Abbreviations: CI, confidence interval; RR, relative risk; total Mg, dietary magnesium plus supplementary magnesium intake; WC, waist circumference
Model 1 was adjusted for age, sex, race, regions and the interaction between age and race
Model 2 was adjusted for covariates in Model 1 plus education, income, smoking, alcohol, and physical activity
Model 3 was adjusted for covariates in Model 2 plus calorie intake, regular aspirin use, calcium intake, and LDL cholesterol For models on dietary Mg intake, supplemental Mg intake was also adjusted.
Sensitivity Analysis: adjusted for covariates in Model 3 plus baseline components of MetS (except baseline WC and glucose as they were not associated with either MetS or Mg in binary analysis).
The median of each quintile was fitted as a continuous variable to compute the Ptrend
As for the associations of dietary and total Mg intake with individual components of MetS, only the association of dietary Mg intake with TG, and the associations of dietary and total Mg intake with WC were statistically significant, Table 3. There was also a borderline inverse association between total Mg intake and high blood glucose and between dietary Mg intake and high HDL.
Table 3:
The RR (95% CI) of the individual components of MetS by quintiles of Mg intake
| Components of MetS | RR (95%CI) | Ptrend | ||||
|---|---|---|---|---|---|---|
| Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 | |||
| Blood Pressure | Total Mg intake | 1.04 (0.97, 1.11) | 0.99 (0.93, 1.07) | 0.95 (0.88, 1.03) | 1.04 (0.94, 1.14) | 0.978 |
| Dietary Mg intake | 0.98 (0.92, 1.05) | 1.03 (0.96, 1.10) | 0.94 (0.87, 1.02) | 1.02 (0.92, 1.13) | 0.901 | |
| Fasting blood glucose | Total Mg intake | 0.85 (0.74, 0.98) | 0.85 (0.73, 0.99) | 0.90 (0.76, 1.05) | 0.86 (0.71, 1.05) | 0.389 |
| Dietary Mg intake | 1.03 (0.89, 1.19) | 1.05 (0.90, 1.22) | 1.03 (0.86, 1.22) | 1.06 (0.86, 1.31) | 0.645 | |
| HDL | Total Mg intake | 1.25 (1.01, 1.55) | 1.04 (0.81, 1.34) | 1.12 (0.86, 1.47) | 0.99 (0.70, 1.40) | 0.720 |
| Dietary Mg intake | 1.07 (0.86, 1.33) | 1.07 (0.85, 1.36) | 0.90 (0.67, 1.20) | 0.87 (0.62, 1.20) | 0.182 | |
| Triglyceride | Total Mg intake | 0.99 (0.87, 1.11) | 1.00 (0.89, 1.14) | 0.92 (0.80, 1.05) | 0.85 (0.72, 1.02) | 0.118 |
| Dietary Mg intake | 0.95 (0.84, 1.07) | 0.91 (0.81, 1.04) | 0.87 (0.76, 1.01) | 0.82 (0.68, 0.98) | 0.023 | |
| WC | Total Mg intake | 0.98 (0.88, 1.09) | 1.00 (0.89, 1.12) | 0.95 (0.84, 1.08) | 0.84 (0.71, 0.99) | 0.121 |
| Dietary Mg intake | 0.89 (0.80, 0.99) | 0.88 (0.79, 0.99) | 0.84 (0.73, 0.95) | 0.71 (0.60, 0.84) | <0.0001 | |
Each outcome was binary based on MetS cutoff points; for blood pressure, either or both systolic (≥130 mmHg) and diastolic (≥ 85 mmHg) blood pressure or blood pressure lowering medication use were used as binary (yes/no).
The values are relative risk and 95% confidence intervals from models adjusted for total calcium intake, age, gender, race, region, education, income, smoking, alcohol intake, regular aspirin use, total calorie intake, physical activity, and baseline values of the components of MetS
Abbreviation: CI, confidence interval; HDL, high-density lipoprotein cholesterol; RR, relative risk; WC, waist circumference
We also looked at the association between Mg intake and the individual components of MetS in continuous form, Supplemental Table S1. There was a trend towards an inverse association between Mg intake and TG in models adjusted for demographic and behavioral variables (model 1 and 2), and a direct association between Mg intake and HDL in models adjusted for demographic variables (model 1), but the association disappeared in fully adjusted models. There was no linear association of Mg intake with FBG, SBP or WC in this study.
An exploratory factor analysis indicated FBG was the weakest component, which was not significantly associated with MetS score, Supplemental Table S2. The path diagrams of the components of MetS, MetS, Mg, and total Mg are presented in Supplemental Figure S4.
Discussion
These results, taken together, showed that Mg intake, particularly from diet, was inversely associated with MetS. The observed associations were mainly driven by the association of Mg intake with TG and WC while there was a borderline inverse association of total Mg with blood glucose and dietary Mg intake with HDL.
The results from our study on the association of dietary Mg with MetS is in agreement with findings from a recent cross-sectional study 26. The results from meta-analyses also indicated similar results for the association of dietary Mg intake with MetS as in our study 10,27. In the meta-analysis by Sarrafzadegan et al. 10, two cohort studies12,13 were included. The Noori et al. 13 study was small (n=160) study among recipients of renal transplants followed up for 1 year. The He et al. study 12 was among 4,637 young adults in the US who were followed-up for 15 years. The Noori et al. study did not find a significant association between Mg intake and MetS but the He et al. study found an inverse association between dietary Mg intake and the risk of MetS. The rest of the studies in the meta-analyses were cross-sectional studies. A recent randomized double blind placebo controlled clinical trial also indicated Mg supplementation improved MetS 28. Among the 198 participants with MetS at baseline and randomly administered 382 mg oral elemental Mg (n=100) or placebo (n=98) daily for 16 weeks, 48% of those in the Mg supplementation group and 77% of those on the placebo group had MetS at the end of follow-up (P = 0.01). The components of MetS were also significantly improved more in the Mg-supplemented group compared to the placebo group 28; there was more reduction in SBP, DBP, glucose, and TG and increase in HDL than the placebo group.
A study using NHANES (1994–1998) cross-sectional data 29 found Mg intake was inversely associated with the risk of MetS among men but not among women. However, in the present cohort study, the observed significant inverse association among the whole study sample disappeared among participants who are black or male and participants from non-Stroke Belt region, presumably due to limited statistical power because of reduced number of participants in those subgroups.
The mechanism through which dietary Mg lowers the risk of MetS may include vasodilation 30 and reduction of BP 31, reduction of chylomicron and blood triglyceride, inhibition of fat absorption 32, promotion of autophosphorylation of insulin receptors by tyrosine kinase thus switching them on and enhancing insulin sensitivity 7, and promotion of glucose transporter protein (GLUT- 4) related to glucose uptake 33. Mg is a cofactor for several rate-limiting enzymes critical for lipid metabolism and it reduces LDL and TG and increase HDL by increasing the activity of lecithin cholesterol acyl transferase 34. Mg intake may also increase lipoprotein lipase activity, which is involved in the conversion of triglycerides to HDL cholesterol 35. The analysis on the association between Mg intake and the individual components of MetS suggests that the observed inverse association in this cohort may be mainly through mechanisms related to TG, central obesity, though there was also a borderline association of Mg intake with FBG and HDL. Path analysis indicated FBG had the weakest association with MetS. Of note, the harmonized definition of MetS emphasizes no single component of the MetS.
The strength of this study includes a large sample size that enabled detection of differences in the risk of MetS between nutrient quintiles. To the best of our knowledge, there has been no previous study that investigated the differences in the association of Mg with MetS by region. The Block 98 FFQ has been validated for Mg intake with a reasonable correlation with the 24-hour dietary record; so, a non-differential misclassification, if existed in this study, might have biased our results towards the null 36. However, a significant inverse association was observed for dietary Mg suggesting that misclassification bias did not affect the result. There are some limitations relevant to the interpretation of the results of our study. In this study, as in previous studies, Mg intake from drinking water was not ascertained; however, region was not an effect modifier in this study. While efforts have been made to include as many potential confounding variables as possible, confounding by unknown or unmeasured variable and residual confounding cannot be completely ruled out. Even though, the Block FFQ had been validated using 24 hours dietary record with a very strong correlation, random measurement error in the Mg intake cannot be 100% ruled out but the use of quintile models generally reduces the effect of such measurement errors as outliers do not make marked impact. Further, missing values in Mg intake or the outcome variables was another limitation of the study. Finally, in the stratified analyses, the statistical power might have been limited which partially explains the disappearance of the association between dietary Mg and MetS in the stratified analysis.
In conclusion, Mg intake was inversely associated with the risk of MetS. These results support the recommendation of foods rich in Mg. Further studies are needed to consolidate the evidence on the observed association and elucidate underlying mechanisms of action.
Supplementary Material
Acknowledgments
Funding source
The study was partially supported by grants to Dr. KH, R01ES021735, R01DK116603, and RF1AG056111. The REGARDS study is supported by a cooperative agreement U01 NS041588 from the National Institute of Neurological Disorders and Stroke, National Institutes of Health, Department of Health and Human Service. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Neurological Disorders and Stroke or the National Institutes of Health. Representatives of the funding agency have been involved in the review of the manuscript but not directly involved in the collection, management, analysis or interpretation of the data. The authors thank the other investigators, the staff, and the participants of the REGARDS study for their valuable contributions. A full list of participating REGARDS investigators and institutions can be found at http://www.regardsstudy.org
Abbreviations:
- BMI
body mass index
- Ca
calcium
- CI
confidence interval
- DBP
diastolic blood pressure
- HDL
high-density lipoprotein
- LDL
low-density lipoprotein
- MetS
metabolic syndrome
- Mg
magnesium
- OR
odds ratio
- REGARDS
REasons for Geographic and Racial Differences in Stroke
- SBP
systolic blood pressure
- TG
triglyceride
- USDA
United States Department of Agriculture
- WC
waist circumference
Footnotes
Conflict of interest
None of the authors declared any conflict of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- 1.Air EL, Kissela BM. Diabetes, the metabolic syndrome, and ischemic stroke: epidemiology and possible mechanisms. Diabetes care 2007;30:3131–40. [DOI] [PubMed] [Google Scholar]
- 2.Aguilar M, Bhuket T, Torres S, Liu B, Wong RJ. Prevalence of the metabolic syndrome in the United States, 2003–2012. JAMA 2015;313:1973–4. [DOI] [PubMed] [Google Scholar]
- 3.Park SH, Kim SK, Bae YJ. Relationship between serum calcium and magnesium concentrations and metabolic syndrome diagnostic components in middle-aged Korean men. Biological trace element research 2012;146:35–41. [DOI] [PubMed] [Google Scholar]
- 4.Kass L, Weekes J, Carpenter L. Effect of magnesium supplementation on blood pressure: a meta-analysis. European journal of clinical nutrition 2012;66:411–8. [DOI] [PubMed] [Google Scholar]
- 5.Jee SH, Miller ER 3rd, Guallar E, Singh VK, Appel LJ, Klag MJ. The effect of magnesium supplementation on blood pressure: a meta-analysis of randomized clinical trials. American journal of hypertension 2002;15:691–6. [DOI] [PubMed] [Google Scholar]
- 6.Floegel A, Stefan N, Yu Z, Mühlenbruch K, Drogan D, et al. Identification of serum metabolites associated with risk of type 2 diabetes using a targeted metabolomic approach. Diabetes 2013;62:639–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Cahill F, Shahidi M, Shea J, Wadden D, Gulliver W, et al. High dietary magnesium intake is associated with low insulin resistance in the Newfoundland population. PloS one 2013;8:e58278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hata A, Doi Y, Ninomiya T, Mukai N, Hirakawa Y, et al. Magnesium intake decreases Type 2 diabetes risk through the improvement of insulin resistance and inflammation: the Hisayama Study. Diabetic medicine : a journal of the British Diabetic Association 2013;30:1487–94. [DOI] [PubMed] [Google Scholar]
- 9.Altura BT, Brust M, Bloom S, Barbour RL, Stempak JG, Altura BM. Magnesium dietary intake modulates blood lipid levels and atherogenesis. Proceedings of the National Academy of Sciences of the United States of America 1990;87:1840–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sarrafzadegan N, Khosravi-Boroujeni H, Lotfizadeh M, Pourmogaddas A, Salehi-Abargouei A. Magnesium status and the metabolic syndrome: A systematic review and meta-analysis. Nutrition (Burbank, Los Angeles County, Calif) 2016;32:409–17. [DOI] [PubMed] [Google Scholar]
- 11.Ju SY, Choi WS, Ock SM, Kim CM, Kim DH. Dietary magnesium intake and metabolic syndrome in the adult population: dose-response meta-analysis and meta-regression. Nutrients 2014;6:6005–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.He K, Liu K, Daviglus ML, Morris SJ, Loria CM, et al. Magnesium intake and incidence of metabolic syndrome among young adults. Circulation 2006;113:1675–82. [DOI] [PubMed] [Google Scholar]
- 13.Noori N, Nafar M, Poorrezagholi F, Ahmadpoor P, Samadian F, et al. Dietary Intakes of Fiber and Magnesium and Incidence of Metabolic Syndrome in First Year After Renal Transplantation . Journal of Renal Nutrition 2010;20:101–11. [DOI] [PubMed] [Google Scholar]
- 14.He K, Song Y, Belin RJ, Chen Y. Magnesium Intake and the Metabolic Syndrome: Epidemiologic Evidence to Date. Journal of the CardioMetabolic Syndrome 2006;1:351–5. [DOI] [PubMed] [Google Scholar]
- 15.Howard VJ, Kleindorfer DO, Judd SE, McClure LA, Safford MM, et al. Disparities in stroke incidence contributing to disparities in stroke mortality. Annals of neurology 2011;69:619–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Shrira I, Christenfeld N, Howard G. Exposure to the US Stroke Buckle as a Risk Factor for Cerebrovascular Mortality . Neuroepidemiology 2008;30:229–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Howard VJ, Cushman M, Pulley L, Gomez CR, Go RC, et al. The reasons for geographic and racial differences in stroke study: objectives and design. Neuroepidemiology 2005;25:135–43. [DOI] [PubMed] [Google Scholar]
- 18.Jensen-Otsu E, Austin GL. Antidepressant Use is Associated with Increased Energy Intake and Similar Levels of Physical Activity. Nutrients 2015;7:9662–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Block G, Hartman AM, Dresser CM, Carroll MD, Gannon J, Gardner L. A data-based approach to diet questionnaire design and testing. American journal of epidemiology 1986;124:453–69. [DOI] [PubMed] [Google Scholar]
- 20.Aaron KJ, Campbell RC, Judd SE, Sanders PW, Muntner P. Association of dietary sodium and potassium intakes with albuminuria in normal-weight, overweight, and obese participants in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) Study. The American journal of clinical nutrition 2011;94:1071–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Newby PK, Noel SE, Grant R, Judd S, Shikany JM, Ard J. Race and region have independent and synergistic effects on dietary intakes in black and white women. Nutrition Journal 2012;11:25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Boucher B, Cotterchio M, Kreiger N, Nadalin V, Block T, Block G. Validity and reliability of the Block98 food-frequency questionnaire in a sample of Canadian women. Public health nutrition 2006;9:84–93. [DOI] [PubMed] [Google Scholar]
- 23.Subar AF, Thompson FE, Kipnis V, Midthune D, Hurwitz P, et al. Comparative validation of the Block, Willett, and National Cancer Institute food frequency questionnaires : the Eating at America’s Table Study. American journal of epidemiology 2001;154:1089–99. [DOI] [PubMed] [Google Scholar]
- 24.Alberti KG, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, et al. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation 2009;120:1640–5. [DOI] [PubMed] [Google Scholar]
- 25.Li R, Hertzmark E, Louie M, Chen L, Spiegelman D. The SAS LGTPHCURV9 Macro. Harvard T.H. Schol of Public Health; 2011. [Google Scholar]
- 26.Moore-Schiltz L, Albert JM, Singer ME, Swain J, Nock NL. Dietary intake of calcium and magnesium and the metabolic syndrome in the National Health and Nutrition Examination (NHANES) 2001–2010 data. The British journal of nutrition 2015;114:924–35. [DOI] [PubMed] [Google Scholar]
- 27.Dibaba DT, Xun P, Fly AD, Yokota K, He K. Dietary magnesium intake and risk of metabolic syndrome: a meta-analysis. Diabetic medicine : a journal of the British Diabetic Association 2014;31:1301–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Rodriguez-Moran M, Simental-Mendia LE, Gamboa-Gomez CI, Guerrero-Romero F. Oral Magnesium Supplementation and Metabolic Syndrome: A Randomized Double-Blind Placebo-Controlled Clinical Trial. Advances in chronic kidney disease 2018;25:261–6. [DOI] [PubMed] [Google Scholar]
- 29.Ford ES, Li C, McGuire LC, Mokdad AH, Liu S. Intake of dietary magnesium and the prevalence of the metabolic syndrome among U.S. adults. Obesity (Silver Spring, Md) 2007;15:1139–46. [DOI] [PubMed] [Google Scholar]
- 30.D’Angelo EK, Singer HA, Rembold CM. Magnesium relaxes arterial smooth muscle by decreasing intracellular Ca2+ without changing intracellular Mg2+. The Journal of clinical investigation 1992;89:1988–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Zhang X, Li Y, Del Gobbo LC, Rosanoff A, Wang J, et al. Effects of Magnesium Supplementation on Blood Pressure: A Meta-Analysis of Randomized Double-Blind Placebo-Controlled Trials. Hypertension (Dallas, Tex : 1979) 2016;68:324–33. [DOI] [PubMed] [Google Scholar]
- 32.Kishimoto Y, Tani M, Uto-Kondo H, Saita E, Iizuka M, et al. Effects of magnesium on postprandial serum lipid responses in healthy human subjects. The British journal of nutrition 2010;103:469–72. [DOI] [PubMed] [Google Scholar]
- 33.Murphy E Mysteries of magnesium homeostasis. Circulation research 2000;86:245–8. [DOI] [PubMed] [Google Scholar]
- 34.Itoh K, Kawasaki T, Nakamura M. The effects of high oral magnesium supplementation on blood pressure, serum lipids and related variables in apparently healthy Japanese subjects. British Journal of Nutrition 1997;78:737–50. [DOI] [PubMed] [Google Scholar]
- 35.Rayssiguier Y, Noe L, Etienne J, Gueux E, Cardot P, Mazur A. Effect of magnesium deficiency on post-heparin lipase activity and tissue lipoprotein lipase in the rat. Lipids 1991;26:182–6. [DOI] [PubMed] [Google Scholar]
- 36.Wacholder S, Hartge P, Lubin JH, Dosemeci M. Non-differential misclassification and bias towards the null: a clarification. Occupational and environmental medicine 1995;52:557–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.