Dietary Magnesium Intake and its Association with Leukocyte Telomere Length in US Middle-aged and Elderly Adults

Background: Magnesium supplementation may extend the life span; however, the biological mechanism is still unknown. Leukocyte telomere length (LTL) is a marker of cell aging and biological health in humans. Data concerning whether magnesium supplementation can maintain telomere length, thus prolonging life are limited. Purpose: To investigate the association between dietary magnesium intake and LTL in US middle-aged and elderly adults. Methods: A total of 4039 US adults aged ≥ 45 years from National Health and Nutrition Examination Survey (NHANES, 1999-2002). Dietary magnesium intake was collected by a trained interviewer using 24-hour dietary recall method and LTL was obtained using the quantitative polymerase chain reaction (PCR) method. Multivariate linear regression analysis was performed to evaluate the the crude and adjusted association of dietary magnesium intake with LTL. Results: The overall mean (SD) of LTL was 5.6 (0.6) kp. After adjusting potential confounders, every 1 mg increase in log-transformed dietary magnesium intake was associated with 0.20 kp [95% condence intervals (CI): 0.05-0.34] longer LTL. Participants with the highest tertile ( ≥ 299 mg) of dietary magnesium intake had statistically signicant longer LTL (β = 0.07, P=0.038) compared with the lowest tertile (< 198 mg), with signicant linear trends across tertiles. Moreover, the association between dietary magnesium intake and LTL was signicantly stronger in participants with higher levels of education ( ≥ high school compared with < high school, P for interaction = 0.002). E-value analysis suggested robustness to unmeasured confounding. Conclusions: Our ndings suggested that increased dietary magnesium intake was associated with longer LTL. Theses warrant additional investigation.


Introduction
Telomeres, the TTAGGG repetitive DNA at the ends of linear chromosomes, are important and active controllers of cellular lifespan and chromosome integrity in eukaryotes cells [1]. Telomere attrition is an integral part of the end replication problem [2]; thus, leukocyte telomere length (LTL) shortening has been viewed as a useful bioindicator for cellular aging [3]. In addition, acceleration of the rate of telomeric sequence loss is a feature of a plethora of adverse health outcomes [4,5]. Shortened LTL been reported to be linked with increased risk for numerous chronic conditions, including cardiovascular disease (CVD) [6][7][8], diabetes mellitus [9,10], alzheimer's disease [11], hypertension [12], and cancer [13]. Growing evidence suggests that LTL can be in uenced by lifestyle factors, such as smoking, physical activity and energy intake [14][15][16]. Recently, the importance of nutritional factors on LTL has been increasingly recognized [17].
Magnesium is an essential element, as a cofactor, by in excess of 300 enzymatic reactions required to maintain homeostasis [18]. Diet is the major source for magnesium in the human. Nuts, seeds, leafy vegetables or whole-grain cereals are well-recognized dietary sources of magnesium [19]. Unbalanced magnesium intake can cause adverse health effects [19,20]. Observational studies have shown that magnesium de ciency is associated with poor cardio-metabolic conditions [20]. An increasing body of epidemiologic evidence reported that higher dietary magnesium intake could exert bene cial effects on CVD risk factors by improving glucose and insulin metabolism, ameliorating lipid pro le and by its actions as an antihypertensive and anti-in ammatory agent [19]. Previous studies have shown that magnesium intake may extending the life span. The process is considered to be associated with the involvement of magnesium in many metabolic processes including ATP-dependent biochemical reactions, synthesis of DNA, RNA expression, cellular excitability and cellular health span [21]. However, whether LTL plays a role in prolonging lifespan with magnesium intake remains unclear. Although some studies have explored the associations between minerals intake (e.g., copper, zinc and selenium) and LTL [22][23][24], the association of dietary magnesium intake with LTL has rarely been examined. Of note, the possible effect modi ers for the dietary magnesium intake-LTL association have not been fully investigated in previous studies.Therefore, the present study aimed to address the knowledge gap by examining the association of dietary magnesium intake with LTL and explore any possible effect modi ers in US middle-aged and elderly adults using a large population based survey data, the National Health and Nutrition Examination Survey (NHANES).

Study design and population
NHANES, conducted by the Centers for Disease Control and Prevention (CDC), was an ongoing repeated cross-sectional study designed to assess the health and nutritional status of adults and children in the United States. The Ethics Review Board of the National Center for Health Statistics (NCHS) approved the NHANES study protocols. Written informed consents were obtained from all study participants. More detailed information is available at http://www.cdc.gov/nchs/nhanes.htm.
The data from NHANES 1999NHANES -2000NHANES and 2001NHANES -2002 were combined for these analyses because LTL was assessed in these two data collection cycles. In total, 7827 participants with LTL data were enrolled.
Considering that LTLwas associated with age-related chronic diseases and middle-aged and elderly people were more prone to malnutrition, we excluded participants aged < 45 years old (n = 3529) and with missing information on dietary magnesium intake (n = 259). Finally, a total of 4039 participants were included in this study (Fig. 1).

Dietary magnesium intake
Dietary intake information was assessed via 24h recall obtained by a trained interview from What We Eat in America survey which was conducted in the Mobile Examination Center. The types and amounts of all foods and beverages during the 24-h period prior to the interview were collected with the use of a computer-assisted dietary interview system. The dietary magnesium intake was estimated based on the University of Texas Food Intake Analysis System and U.S. Department of Agriculture Survey Nutrients Database [22]. The nutrient estimates did not include nutrients obtained from medications or dietary supplements.

LTL assessment
The detail methods for LTL quanti cation have been described previously [12,23,25]. The telomere length relative to standard reference DNA (T/S ratio) was obtained in the blood leukocytes using real-time quantitative polymerase chain reaction (qPCR) [26]. Each sample was assigned to duplicate wells in a 96 well plate and assayed three times on three different days. Each panel contained eight control DNA samples to normalize the variability between the two tests. The inter-assay coe cient of variation was 6.5%. The formula of the conversion from T/S ratio to kp is [3274 + 2413 × (T/S)]/1000 [25]. The conversion from T/S ratio to kp was calculated based on a comparison of telomeric restriction fragment length from Southern blot analysis and T/S ratios using DNA samples from the human diploid broblast cell line IMR90 at different population doublings.

Covariates
The following variables were used to construct the fully adjusted model. Continuous variables included age (years), body mass index (BMI, kg/m 2 ), poverty to income ratio (PIR), sample weight, alcohol consumption (gm), total cholesterol (TC, mg/dL), high density lipoprotein cholesterol (HDL-C, mg/dL), triglycerides (TG, mg/dL), fasting blood glucose (FBG, mg/dL), dietary ber (g) and total energy intake (kcal). Categorical variables consisted of sex (male, female), race/ethnicity (non-Hispanic white, non-Hispanic black, Mexican American, other Hispanic or other), education (less than high school, high school, or greater than high school), smoking (never smoker, former smoker or current smoker), physical activity (sedentary, low, moderate, or high), self-reported history of diabetes and hypertension.
Hypertension was de ned as having a history of hypertension, a systolic blood pressure (SBP) ≥ 140 mmHg and/or diastolic blood pressure (DBP) ≥ 90 mmHg, and using antihypertensive medications [12,27].

Statistical Analysis and Sensitivity analysis
Sample weights were used for analyses to account for the complex survey design and non-response of NHANES [28]. Weighted means, proportions and standard error (Se) were calculated for baseline characteristics using survey sample weights. A weighted linear regression model (continuous variables) or weighted chi-square test (categorical variables) were used to calculate for differences among different dietary magnesium intake groups (tertiles). Because the distribution of values for dietary magnesium intake was strongly skewed toward the upper end, the dietary magnesium intake was log-transformed to better approximate a normal distribution. We applied multivariate linear regression analysis to evaluate the independent association between dietary magnesium intake and LTL. We constructed two adjusted models: Model 1 was adjusted for age, sex, race, education status, smoking status, alcohol consumption and PIR; model 2 was further adjusted for physical activity, BMI, hypertension, diabetes, TC, TG, HDL-C, FBG, sample weight, dietary ber and total energy intake. Results were presented in coe cients (β) with the corresponding 95% con dence intervals (CIs). Multivariate linear analysis with cubic spline functions model and smooth curve tting (penalized spline method) was further conducted to characterize the shape of the relationship between dietary magnesium intake and LTL [29].
To ensure the robustness of data analysis, we also did the sensitivity analyses. We performed testing for linear trends by entering the median value of each category of dietary magnesium intake as a continuous variable in the models. Moreover, we explored the potential for unmeasured confounding between baPWV and rst stroke by calculating E-values. The E-value quanti es the required magnitude of an unmeasured confounder that couldnegate the observed association between dietary magnesium intake and LTL. In addition, possible modi cations of the association between dietary magnesium intake and LTL were also assessed for the following variables: sex (females vs. males), age (< 65 vs. ≥65 years), BMI (< 30 vs. ≥30 kg/m 2 ), current smoking (yes vs. no), education (< high school vs. ≥high school), diabtese (yes vs. no) and hypertension (yes vs. no).
All P values were two-sided with a signi cance level of < 0.05. The analyses were performed using the statistical package R (http://www.R-project.org, The R Foundation).

Baseline Characteristics of Study Participants
A total of 4039 study participants aged 45-85 years (weighted mean age: 15.5 ± 2.3 years; 47.4% men) were included in this nal data analysis. As shown in Fig. 2, the mean intake of magnesium was 270.7 ± 134.8 mg (median 246.0 mg). The average level of LTL was 5.6 ± 0.6 kp (median 5.5 kp). The distributions of dietary magnesium intake and LTL were skewed toward the upper end. The weighted distributions of 4039 participants' sociodemographic characteristics and other covariates according to dietary magnesium intake tertiles were shown in Table 1. The ranges of dietary magnesium intake for tertiles 1-3 were < 198, 198-299, and ≥ 299 mg, respectively. Compared with tertile 1 and tertile 2 of dietary magnesium intake, participants in tertile 3 seemed to be younger, to be more males and nonhispanic whites, to have higher educational levels and physical activity levels, to have lower rate of current smoking and hypertension, to have higher values in PIR, alcohol intake, dietary ber, total energy intake and LTL and to have lower values in BMI and HDL-C (all P < 0.01).  Table 2 showed the association between dietary magnesium intake and LTL using multivariate linear regression analyses. In the crude model, log-transformed dietary magnesium intake, as a continuous variable, was positively associated with longer LTL (β = 0.11, 95% CI: 0.03-0.20, P = 0.009). The association between dietary magnesium intake and LTL was still observed after adjustment for different potential confounders (all P < 0.01). In the fully adjusted model (Model ), every 1 mg increase in logtransformed dietary magnesium intake was associated with 0.20 kp (95%CI: 0.05-0.34) longer LTL. We also converted dietary magnesium intake from a continuous variable to a categorical variable (tertiles). Compared with the tertile 1 of dietary magnesium intake (< 198 mg), the participants in the highest tertile (dietary magnesium intake ≥ 299 mg) had statistically signi cant longer LTL (β = 0.07, 95% CI: 0.01-0.13). Tests for linear trend (P for trend) were signi cant and consistent with the P value when dietary magnesium intake was used as a continuous variable, suggesting a signi cant linear trend among tertiles of dietary magnesium intake and LTL. We further explored the shape of the dose-response relationship by using GAM models. As shown in Fig. 3, the association of dietary magnesium intake with LTL was linear.

Sensitivity Analysis
We generated an E-value to assess the sensitivity to unmeasured confounding. Dietary magnesium intake was positively associated with an increase in rst stroke by multivariable analysis (β = 0.20, 95% CI: 0.05-0.34). The E-value was 2.05, meaning that residual confounding could explain the observed association if there exists an unmeasured covariate having a relative risk association ≥ 2.05 with both dietary magnesium intake and LTL. Therefore, it is unlikely that an unmeasured or unknown confounder would have a substantially greater effect on LTL than these known risk factors. Consistently, the primary ndings were robust.
In addition, we further explored the role of covariables on the association between dietary magnesium intake and LTL. As shown in Fig. 4, the effect of log-transformed dietary magnesium intake on LTL was signi cantly stronger in participants with higher levels of education (< high school: β = 0.11, 95%CI: -0.11-0.34; ≥ high school: β = 0.27, 95%CI: 0.08-0.47, P for interaction = 0.002). None of the other variables, including sex, age, BMI, current smoking, diabtese and hypertension, showed signi cant effect modi cation on the association between dietary magnesium intake and LTL (all P for interaction > 0.05).

Discussion
In this study, dietary magnesium intake was independently and positively associated with longer LTL in US middle-aged and elderly participants, even after adjusting for covariates. Furthermore, this association was more pronounced in populations with higher levels of education. Our ndings suggested magnesium might play an important role in LTL.
We focused on dietary magnesium and its association with LTL because diet represented an easily modi able intervention target. Prior reviews indicated that magnesium might have protective effects on telomere attrition [30] and correcting magnesium de ciencies might prolong life [31]. However, whether LTL plays a role in prolonging lifespan with magnesium intake remains unclear. Of note, only two previous studies have reported the association between magnesium and LTL with inconsistent results. Yu Jie et al [32]. conducted a cross-sectional study of 467 participants with a mean magnesium intake of 247.13 mg/day and found an inverse relationship between dietary magnesium and LTL. Nathan J.
O'Callaghan et al [33]. used data from 23 South Australian aged 65 years or older and found that a negative association between telomere length and serum magnesium levels (r=-0.61, P = 0.001 respectively). The previous studies have many limitations, such as small sample size and simple statistical methods. They had thus far only focused on exploring the correlation, not multivariate regression analysis. Besides, none of them discussed the possibility of a nonlinear relationship. Furthermore, whether this association was modi ed by some risk factors of LTL needs to be veri ed.
Magnesium is the second most abundant cation in cellular systems and plays a role in numerous biological functions including maintenance of DNA structure, thus it is di cult to reconcile the role of high magnesium in telomere shortening.
In our study, we provided some new insights into this eld. First, we observed that dietary magnesium intake was positively associated with longer LTL. Also, our study con rmed a signi cant linear relationship between dietary magnesium intake and LTL using smooth curve tting. The ndings suggested that magnesium were conducive to the longer life expectancy which might have an intimate connection with LTL. Several possible mechanisms for this association including (1) Magnesium plays a role in telomere maintenance and the activity of telomerase; thus, increasing magnesium supplement could extend LTL [31]. (2) Long-term intake of magnesium have demonstrated anti-in ammatory and antioxidative properties which may impact on LTL [34,35]. However, more studies are needed to con rm our results and further examine the underlying mechanisms. Second, subgroup analyses showed that education level was a signi cant modi er: a stronger association was found in participants with higher levels of education (≥ high school). The biological mechanism underlying high levels of education × high dietary magnesium intake is still unknown. In our current study, participants with the higher levels of education had longer LTL; however, among those with lower education level, the LTL was shorten. Some studies reported that excess sedentary behavior and lack of physical activity could contribute to shortened telomeres [36,37]. A plausible biological explanation for the interaction may be due to the fact that poor lifestyle, behavior or environmental conditions related to lower education level has an indirect in uence on the association between dietary magnesium intake and LTL. Moreover, one study demonstrated that low educational attainment might be an indicator of long-term socioeconomic status trajectories, and be associated with accumulated allostatic load resulting in telomere shortening [38].Taken together, further research is needed to examine the association between education level, dietary magnesium intake and LTL.
The mean level of dietary magnesium intake as seen in our study was 270.7 mg/day, which was less than the general recommendations for 300 mg magnesium intake a day. Although the serum magnesium was also unavailable, the data from NHANES 2006 showed that low serum magnesium concentrations were 36.3% and 31.0% for females and males, respectively [39]. Moreover, epidemiological studies reported that dietary magnesium intake was inadequate in the US population as well as in other populations [40,41]. 68% of Americans [40] and 72% of middle aged French adults have been shown to consume below the recommended intake of magnesium. These ndings suggested that dietary magnesium intake was inadequate in US populations. Thus, increasing the dietary magnesium intake in the US population may be an important public health goal.
Our study has some strengths. First, this study was the rst report to explore the association between dietary magnesium intake and LTL in US middle-aged and elderly adults. Second, the study populations were randomly invited to NHANES which applied rigorous quality controls to the procedures. Third, we provided adequate statistical rationale to evaluate the association between dietary magnesium intake and LTL, a feature that was lacking in previous studies. Some limitations of our study should be also noted. First, as a cross-sectional design, it had less power to infer the causal association between dietary magnesium intake and LTL. Further prospective cohort studies are needed to verify these ndings. Second, as in all observational studies, even though known potential confounding factors were controlled for, there might have been still uncontrolled confounding due to unmeasured differences in behaviors or other factors. We used the E-value sensitivity analysis to quantify the potential implications of unmeasured confounders and found that an unmeasured confounder was unlikely to explain the entirety of dietary magnesium intake effect. Third, fruit, vegetable, and whole grain intake were not adjusted.
However, the dietary magnesium intake was estimated based on the types and amounts of all foods and beverages. Fourth, the dietary intake was assessed by 24-h dietary recall interviews which might lead to imprecise estimates due to day-to-day variations in diet; accounting for the association between blood magnesium levels and LTL would be an important contribution in the future.

Conclusions
In summary, this study demonstrated that dietary magnesium intake was independently and positively associated with LTL in US middle-aged and elderly participants, and this association was more pronounced in populations with higher levels of education. Theses warrant additional investigation.

Declarations
Ethics approval: NHANES study protocols were approved by the research ethics review board of the National Center for Health Statistics.
Consent to participate: Written informed consent was acquired from each participant Consent for publication: Not applicable.
Availability of data and materials: The datasets are available on DataDryad (https://doi.org/10.

5061/dryad.d5h62).
Competing interests: The authors declare that they have no con ict of interest.  Flow chart of the study participants.

Figure 2
The distributions of dietary magnesium intake and leukocyte telomere length.

Figure 3
Dose-response relationship between dietary magnesium intake and leukocyte telomere length. Adjusted for age, sex, BMI, race, education status, smoking status, alcohol consumption, physical activity, PIR, hypertension, diabetes, TC, TG, HDL-C, FBG, sample weight, dietary ber and total energy intake.