Serum Magnesium Levels Are Negatively Associated with Obesity and Abdominal Obesity in Type 2 Diabetes Mellitus: A Real-World Study

Article information

Diabetes Metab J. 2024;.dmj.2023.0401
Publication date (electronic) : 2024 May 29
doi : https://doi.org/10.4093/dmj.2023.0401
1Department of Endocrinology and Metabolism, Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai Clinical Center for Diabetes, Shanghai Diabetes Institute, Shanghai Key Laboratory of Diabetes Mellitus, Shanghai Key Clinical Center for Metabolic Disease, Shanghai, China
2Department of Endocrinology, Eastern Theater Air Force Hospital of People’s Liberation Army, Nanjing, China
3Clinical Research Center, Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China
Corresponding author: Lian-Xi Li Department of Endocrinology and Metabolism, Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai Clinical Center for Diabetes, Shanghai Diabetes Institute, Shanghai Key Laboratory of Diabetes Mellitus, Shanghai Key Clinical Center for Metabolic Disease, 600 Yishan Road, Shanghai 200233, China E-mail: lilx@sjtu.edu.cn
*Man-Rong Xu and Ai-Ping Wang contributed equally to this study as first authors.
Received 2023 November 8; Accepted 2024 March 6.

Abstract

Background

There remains controversy over the relationship between serum magnesium levels and obesity in type 2 diabetes mellitus (T2DM). Therefore, the aim of this study was to assess whether there is any association of serum magnesium levels with obesity and abdominal obesity in T2DM.

Methods

This cross-sectional, real-world study was conducted in 8,010 patients with T2DM, which were stratified into quintiles according to serum magnesium levels. The clinical characteristics and the prevalence of obesity and abdominal obesity were compared across serum magnesium quintiles in T2DM. Regression analyses were used to evaluate the relationship of serum magnesium with obesity and abdominal obesity in T2DM (clinical trial registration number: ChiCTR1800015893).

Results

After adjustment for age, sex, and duration of diabetes, the prevalence of obesity and abdominal obesity was significantly declined across magnesium quintiles (obesity: 51.3%, 50.8%, 48.9%, 45.3%, and 43.8%, respectively, P<0.001 for trend; abdominal obesity: 71.5%, 70.5%, 68.2%, 66.4%, and 64.5%, respectively, P=0.001 for trend). After controlling for confounders, there were clearly negative associations of serum magnesium levels and quintiles with obesity and abdominal obesity in T2DM. Moreover, C-reactive protein partly mediates the effect of serum magnesium on obesity and abdominal obesity (P=0.016 and P=0.004, respectively).

Conclusion

The significantly negative relationship between serum magnesium and the risk of obesity and abdominal obesity was observed in T2DM. Furthermore, the independently negative association of serum magnesium with obesity may be explained by its anti-inflammatory functions. Serum magnesium levels may be applied to assess the risk of obesity and abdominal obesity in T2DM.

KEY FIGURE

Highlights

· Low serum magnesium levels increased the risk of obesity and abdominal obesity in T2DM.

· The negative link between serum magnesium and obesity may be due to its anti-inflammatory effects.

· Serum magnesium levels can help assess the risk of obesity and abdominal obesity in T2DM.

INTRODUCTION

Magnesium ions have been identified as essential cofactors in more than 600 enzymatic reactions in the body, which regulate a variety of physiological activities such as protein synthesis, glycolysis, and energy metabolism [1]. In addition, increasing number of studies have shown that the decrease in extracellular magnesium ions leads to the increase in intracellular calcium and iron ions, which induce increased production and decreased clearance of peroxides and trigger inflammatory pathways [2-4]. Therefore, the change of magnesium ions can lead to pathological changes such as endothelial cell destruction and insulin resistance, and is also associated with chronic inflammatory diseases such as atherosclerosis and chronic kidney disease (CKD) [2,4,5]. Interestingly, in a rat model with metabolic syndrome (MetS) combined with CKD, supplementation with dietary magnesium significantly reduces oxidative stress and inflammatory responses, thereby acting as an intravascular protector [6].

Due to the vital physiological role of magnesium ions, multiple clinical studies have observed a significant correlation between serum magnesium levels and a range of metabolic diseases such as diabetes and cardiovascular disease [7-10]. For example, Zhang et al. [11] recently reported that low serum magnesium not only was a risk factor for the development of diabetes, but also significantly and positively correlated with diabetes-related indicators such as fasting plasma glucose (FPG), postprandial blood glucose, and glycosylated hemoglobin (HbA1c), suggesting that hypomagnesemia (serum magnesium <0.75 mmol/L) can lead to insulin resistance and poor glycemic control.

As one of the metabolic disorders, obesity is also a key risk factor for other metabolic diseases such as diabetes, non-alcoholic fatty liver disease, and cardiovascular disease [12,13]. Furthermore, there have been increasing evidences that abdominal obesity may have more serious detrimental effects on metabolic homeostasis [14]. Given the close association of serum magnesium levels with metabolic diseases, a wide variety of studies have observed a strong relationship between serum magnesium and obesity. For example, Hassan et al. [15] revealed that serum magnesium levels were significantly higher in overweight and obese children than in children with normal weight, and negatively associated with body mass index (BMI).

However, another previous study conducted with a large sample of South Asians and white Canadians found that there was no difference in serum magnesium levels among different BMI groups in men [16]. Similarly, Alasfar et al. [17] also observed no obvious differences in serum magnesium levels between non-obese and morbidly obese women. Therefore, the association between serum magnesium and obesity in different populations is controversial and needs further clarification.

Furthermore, the correlation between serum magnesium and obesity also remains a controversy in type 2 diabetes mellitus (T2DM) population [17-19]. Chu et al. [18] found that BMI decreased significantly with increasing serum magnesium tertiles in patients with T2DM. Additionally, another study involving 210 elderly patients with T2DM demonstrated that high magnesium intake reduced the incidence of centripetal obesity, high body fat ratio and high BMI, but no correlation was observed between serum magnesium and metabolic parameters [20]. In contrast, no significant relationship between serum magnesium and BMI was observed in an investigation with 2,222 patients with T2DM [19]. Moreover, relevant studies about the relationship between serum magnesium and obesity as well as abdominal obesity remains sparse in Chinese patients with T2DM. Therefore, the aim of our study was to explore the actual association of serum magnesium with obesity, as well as abdominal obesity in Chinese T2DM patients with relatively large samples in a real-world setting, respectively.

METHODS

Study population

A total of 11,805 patients with T2DM were enrolled in the present cross-sectional, real-world study. These participants were consecutively recruited from the Department of Endocrinology and Metabolism, Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine from January 2003 to December 2012. Among them, these subjects were excluded as follows: (1) age below 18 years old; (2) incomplete physical examination and laboratory test data; (3) acute diabetic complications such as diabetic ketoacidosis; (4) taking magnesium supplements or taking drugs that influences serum magnesium levels; (5) systemic inflammatory disease; and (6) progressive malignancy. Ultimately, 8,010 T2DM subjects were included in the final analysis.

All participants were interviewed to obtain their information on the duration of diabetes (DD), smoking habits, alcohol consumption, history of hypertension (HTN), and use of medications, including sulfonylureas (SUs), metformin, and insulin or insulin analogues (IIAs). This study was guided by the principles of the Declaration of Helsinki and approved by the Ethics Committee of the Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (approval number: 2018-KY-018(K)). Each subject signed a written consent form.

Physical examination and laboratory measurements

The physical examination data were obtained according to the standard protocol, including systolic blood pressure (SBP), diastolic blood pressure (DBP), body weight, height, waist circumference (WC), and hip circumference. BMI, waist-to-hip ratio (WHR), and waist-to-height ratio (WHtR) were calculated according to our previous method [21-24]. Venous blood samples were collected from all subjects after an overnight fast and 2 hours after breakfast. Laboratory measurements included FPG, 2-hour postprandial plasma glucose (2hr PPG), fasting C-peptide (FCP), 2-hour postprandial C-peptide (2hr CP), HbA1c, triglyceride (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), alanine aminotransferase (ALT), creatinine (Cr), serum uric acid (SUA), urinary albumin excretion (UAE), and C-reactive protein (CRP). Serum magnesium levels were determined using the xylidyl blue method by the Hitachi 7600 analyzer (Hitachi, Tokyo, Japan) [18]. The 24-hour UAE, estimated glomerular filtration rate (eGFR), and the updated homoeostasis model assessment of insulin sensitivity/insulin resistance were calculated, as described in our previous study [21].

Diagnostic criteria

Normal serum magnesium levels were defined as 0.65 to 1.05 μmol/L according to the normal reference range of serum magnesium in our hospital. Based on normal serum magnesium levels, low serum magnesium was defined as serum magnesium levels <0.65 μmol/L and high serum magnesium as >1.05 μmol/L. The definitions of smoking, alcohol consumption, and HTN were consistent with our previous studies [25]. With reference to our previous studies [26], T2DM was diagnosed based on the 1999 World Health Organization criteria. According to our previous description, obesity was defined as a BMI ≥25 kg/m2, and the diagnostic criteria for the severity of obesity was defined as BMI 25 to 30 kg/m2 for mild obesity, 30 to 35 kg/m2 for moderate obesity, and over 35 kg/m2 for severe obesity [27]. According to the Chinese criteria proposed by the Chinese Obesity Task Force and our previous study, abdominal obesity was defined as a WC ≥90 cm in men and ≥80 cm in women [13]. In addition, mild abdominal obesity was specified as a WC of 90 to 100 cm for men and 80 to 90 cm for women, moderate abdominal obesity 100 to 110 cm for men and 90 to 100 cm for women, severe abdominal obesity ≥110 cm for men and ≥100 cm for women, respectively [13].

Statistical analyses

We analyzed all data using SPSS version 15.0 software (SPSS Inc., Chicago, IL, USA) and R version 4.0.2 software (R Foundation for Statistical Computing, Vienna, Austria). Normally and non-normally distributed data were expressed as mean±standard deviation and median with interquartile range, respectively. One-way analysis of variance (ANOVA) with least significant difference was carried out to determine group differences in normally distributed variables, whereas the Kruskal-Wallis test was performed to compare non-normally distributed variables. Prevalence data and frequency differences were compared using chi-square tests. When controlling for confounding factors, binary logistic regression and general linear regression were conducted to assess the differences between groups.

Restricted cubic splines were used with three knots at the 10th, 50th, and 90th centiles to flexibly model the relationship between serum magnesium and obesity/abnormal obesity after adjusting covariates in model 6. Moreover, we conducted a causal mediation analysis to assess whether CRP could be a potential mediator linking serum magnesium to obesity/abnormal obesity, using the R “mediation” package. P<0.05 was considered statistically significant.

To assess the associations of serum magnesium levels and quintiles with obesity and abdominal obesity in T2DM, six logistic regression models were constructed: model 1 was unadjusted; model 2 was adjusted for age, gender, and DD; model 3 was additional adjusted for medical history, including HTN, smoking, and alcohol drinking; model 4 was additional adjusted for the use of drugs, including the use of IIAs, metformin, and SUs; model 5 included additional adjustments for physical examination data, including SBP and DBP; and model 6 included further adjustments for laboratory results, including ALT, TC, HDL-C, LDL-C, TG, eGFR, Cr, SUA, UAE, FPG, 2hr PPG, HbA1C, FCP, 2hr CP, and CRP.

RESULTS

Characteristics of the subjects according to serum magnesium quintiles

Table 1 shows the clinical characteristics of the T2DM patients. According to serum magnesium levels, the T2DM patients were divided into serum magnesium quintiles with cut-offs of <0.82, 0.82–0.87, 0.88–0.91, 0.92–0.96, and >0.96 mmol/L. Marked differences were observed for sex and age across serum magnesium quintiles (P<0.05). After controlling for age and sex, significantly elevated trends in 2hr CP and Cr were observed with increasing quintiles of serum magnesium (all P<0.05). Conversely, the use of IIAs and metformin, DD, FPG, 2hr PPG, HbA1c, UAE, and eGFR were markedly declined across serum magnesium quintiles (all P<0.05). Additionally, after adjustment for age and sex, the prevalence of HTN differed significantly among the serum magnesium quintiles, but without a trend towards an increase or decrease (P<0.05).

Characteristics of the subjects according to serum magnesium quintiles

Comparisons of serum magnesium levels in different groups

Fig. 1 compares serum magnesium levels in different groups in T2DM subjects. Serum magnesium levels were significantly lower in men than in women after controlling for age and DD (P<0.001, Fig. 1A). Furthermore, there was a noticeably increasing trend in serum magnesium levels with advancing age, while an inverse trend was observed with prolonged DD (P< 0.001 for trend, Fig. 1D; P=0.004, Fig. 1E). Additionally, serum magnesium levels demonstrated a clear decline with increasing severity of obesity and abdominal obesity. Notably, the patients with obesity or abdominal obesity exhibited substantially lower serum magnesium levels compared to those without (P<0.001 for trend, Fig. 1B and C).

Fig. 1.

Comparisons of serum magnesium levels in different groups. (A) Comparison of serum magnesium levels between men and women with type 2 diabetes mellitus after adjusting for age and duration of diabetes (DD) (P<0.001). (B) Comparison of serum magnesium levels stratified by the degree of obesity after adjusting for age, sex, and DD (P<0.001 for trend). (C) Comparison of serum magnesium levels stratified by the degree of abdominal obesity after adjusting for age, sex, and DD (P<0.001 for trend). (D) Comparison of serum magnesium levels stratified by age after adjusting for DD and sex (P<0.001 for trend). (E) Comparison of serum magnesium levels stratified by DD after adjusting for age and sex (P<0.001 for trend).

Comparisons of the prevalence of obesity and abdominal obesity stratified by sex, age, and DD

Supplementary Fig. 1 compares the prevalence of obesity and abdominal obesity stratified by sex, age, and DD in T2DM patients. There was no significant difference in the prevalence of obesity between genders (48.8% in men and 47.4% in women, P=0.509, Supplementary Fig. 1A) after adjustment for age and DD; however, the prevalence of abdominal obesity was markedly higher in women (81.8%) than in men (57.4%) (P<0.001, Supplementary Fig. 1A). In addition, the prevalence of obesity significantly decreased with increasing age, while the prevalence of abdominal obesity obviously increased with advancing age (P<0.001 for trend in obesity and P=0.002 for trend in abdominal obesity, respectively; Supplementary Fig. 1B). Furthermore, statistically significant differences were found in the prevalence of obesity and abdominal obesity among different DD groups (P=0.013 for trend in obesity and P=0.019 for trend in abdominal obesity, respectively; Supplementary Fig. 1C).

Comparisons of the prevalence of obesity and abdominal obesity among serum magnesium quintiles

Fig. 2 compares the prevalence of obesity and abdominal obesity among serum magnesium quintiles in patients with T2DM. After controlling for age, sex, and DD, a clear downward trend in the prevalence of obesity and abdominal obesity was observed with increasing quintiles of serum magnesium (51.3%, 50.8%, 48.9%, 45.3%, 43.8% for each quintile in obesity, respectively, P<0.001 for trend; and 71.5%, 70.5%, 68.2%, 66.4%, 64.5% for each quintile in abdominal obesity, respectively, P=0.001 for trend) (Fig. 2A). Interestingly, the prevalence of obesity and abdominal obesity gradually declined from low to high serum magnesium group (P=0.030 for trend in obesity, and P=0.088 for trend in abdominal obesity, respectively; Fig. 2B). Additionally, the prevalence of severe obesity and abdominal obesity significantly increased with the decreased serum magnesium quintile (P<0.001 for trend in obesity, Fig. 2C; P<0.001 for trend in abdominal obesity, Fig. 2D).

Fig. 2.

Comparisons of the prevalence of obesity and abdominal obesity among serum magnesium quintiles. (A) Comparison of the prevalence of obesity and abdominal obesity stratified by serum magnesium quintiles after adjusting for age, sex, and duration of diabetes (DD) (P<0.001 for trend in obesity, P=0.001 for trend in abdominal obesity). (B) Comparison of the prevalence of obesity and abdominal obesity in low, normal, high serum magnesium group after adjusting for age, sex, and DD (P=0.030 for trend in obesity, P=0.088 for trend in abdominal obesity). (C) Comparison of the prevalence of mild, moderate, and severe obesity across serum magnesium quintiles after adjusting for age, sex, and DD (P<0.001 for trend). (D) Comparison of the prevalence of mild, moderate, and severe abdominal obesity across serum magnesium quintiles after adjusting for age, sex, and DD (P<0.001 for trend).

Linear relationship between serum magnesium levels and BMI, WC, WHR, and WHtR

In Supplementary Fig. 2, correlation analysis displays the linear relationship between serum magnesium concentration and anthropometric measures including BMI, WC, WHR, and WHtR. After controlling for age, sex, DD, a weakly but significantly negative correlation was found between serum magnesium levels and BMI (R=–0.077, P<0.001, Supplementary Fig. 2A), WC (R=–0.074, P<0.001, Supplementary Fig. 2B), WHR (R=–0.089, P<0.001, Supplementary Fig. 2C), and WHtR (R=–0.078, P<0.001, Supplementary Fig. 2D) in T2DM patients.

Association of serum magnesium levels with obesity and abdominal obesity

Table 2 analyses the associations of serum magnesium levels with obesity and abdominal obesity in T2DM subjects. The obviously negative associations of serum magnesium levels with obesity and abdominal obesity were observed (model 1) (odds ratio [OR], 0.879; 95% confidence interval [CI], 0.841 to 0.919; P<0.001 for trend in obesity, and OR, 0.881; 95% CI, 0.840 to 0.923; P<0.001 for trend in abdominal obesity, respectively). Furthermore, significant correlations between serum magnesium levels and obesity/abdominal obesity remained even after further adjustment for various confounding factors in model 2–6 (P<0.001 for trend in obesity in model 2–4 and models 6 respectively; P=0.001 for trend in obesity in models 5; P<0.001 for trend in abdominal obesity in models 2–3; P=0.004 for trend in abdominal obesity in models 4–6).

The association of serum magnesium levels with obesity and abdominal obesity

Association of serum magnesium with obesity and abdominal obesity analyzed by restrictive cubic spline

Supplementary Fig. 3 shows the association of serum magnesium with obesity and abdominal obesity analyzed by restrictive cubic spline (RCS). In RCS, after adequate correction for confounders according to model 6, the OR for obesity were significantly decreased with increasing serum magnesium (P=0.001, Supplementary Fig. 3A). Additionally, when serum magnesium >0.89 mmol/L, the patients had an obviously lower OR value (<1.00) in obesity (P=0.001, Supplementary Fig. 3A). Moreover, significantly lower OR (<1.00) was also observed when serum magnesium levels were higher than 0.89 mmol/L in abdominal obesity (P=0.045, Supplementary Fig. 3B).

Association of serum magnesium quintiles with obesity and abdominal obesity

Table 3 exhibits the associations of serum magnesium quintiles with obesity and abdominal obesity in T2DM subjects. After controlling for multiple confounders, as shown in models 1–6, the prevalence of obesity significantly decreased with increasing serum magnesium quintiles (P<0.001 in models 1 and 2, P=0.001 in model 3, P=0.020 in model 4, P=0.023 in model 5, and P=0.001 in model 6 for trend in obesity, respectively). Furthermore, the prevalence of abdominal obesity also noticeably decreased across the serum magnesium quintiles in patients with T2DM after fully controlling for confounding variables in model 6 (P=0.031 in model 6 for trend in abdominal obesity).

Association of serum magnesium quintiles with obesity and abdominal obesity

Role of CRP in the relationship between serum magnesium and obesity/abdominal obesity

Fig. 3 illustrates the comparisons of serum CRP levels in different groups and the mediation effect of CRP in the association of serum magnesium with obesity and abdominal obesity in patients with T2DM. Classical acute inflammation leads to a marked increase in serum CRP levels. In contrast, the hallmark of low-grade chronic inflammation is a slight rise in CRP being a range of concentrations of 3 to 10 mg/L [28]. Therefore, to eliminate the effect of acute inflammation such as infection, we excluded the patients with CRP >10 mg/L when analyzing the effect of CRP on the association between serum magnesium and obesity as well as abdominal obesity.

Fig. 3.

Role of C-reactive protein (CRP) in the relationship between serum magnesium and obesity/abdominal obesity. (A) Comparison of serum CRP levels between the patients with and without obesity/abdominal obesity after controlling for gender, age, and duration of diabetes (DD) (P<0.001 for trend in obesity and abdominal obesity). (B) Comparison of serum CRP levels in mild, moderate, and severe obesity/abdominal obesity after controlling for gender, age, and DD (P=0.025 for trend in obesity, P<0.001 for trend in abdominal obesity). (C) Comparison of serum CRP levels across serum magnesium quintiles after adjusting for gender, age, and DD (P=0.046 for trend). (D) The mediation effects of CRP on the association between serum magnesium and obesity (P=0.016). (E) The mediation effects of CRP on the association between serum magnesium and abdominal obesity (P=0.004). ACME, average causal mediation effect; ADE, average direct effects.

We found that CRP was significantly higher in the T2DM patients with obesity/abdominal obesity than in those without obesity/abdominal obesity (P<0.001 for all trend, Fig. 3A). In addition, CRP levels were obviously increased with the degree of severity of obesity and abdominal obesity (P=0.025 for all trends in obesity, P<0.001 for trend in abdominal obesity, Fig. 3B). Moreover, CRP levels showed a significantly negative correlation with serum magnesium quintiles (P=0.046 for trend, Fig. 3C). Additionally, mediation analysis indicated that the average direct effects (ADE) of serum magnesium on obesity was –0.312 (P<0.001), where the average causal mediation effects (ACMEs) of CRP as a mediator was –0.026 (P=0.016), and 7.54% of the total effect of serum magnesium on obesity could be explained by the ACME of CRP (Fig. 3D). Similarly, serum magnesium on abdominal obesity had an ADE=–0.203 (P<0.001), where the ACME of CRP was –0.024 (P=0.004), accounting for 10.42% of the total effect of serum magnesium on abdominal obesity (Fig. 3E).

DISCUSSION

In this real-world study with relatively large samples, a significantly negative associations of serum magnesium levels with obesity and abdominal obesity were observed in patients with T2DM. Therefore, our findings suggested that low levels of serum magnesium may be an independent risk factor for the occurrence of obesity and abdominal obesity in patients with T2DM.

The present study revealed that serum magnesium levels in patients with T2DM were noticeably higher in men than in women; increased with age and decreased with DD. Consistent with our findings A previous study demonstrated that serum magnesium levels are negatively correlated with estrogen in women [29], which might explain the significantly lower serum magnesium levels in women compared to men in our present study. However, Masood et al. [30] found no significant differences in serum magnesium levels stratified by gender, age, and DD in patients with T2DM. Variations in serum magnesium levels across different study based on age, sex, and DD may be attributed to sample size, ethnicity, and geographical region.

More importantly, previous studies have demonstrated that serum magnesium plays a crucial protective role in metabolic diseases including obesity [15,31,32]. For example, as serum magnesium quartiles ranged from high to low, Shi and Masood [31] observed an increased prevalence of obesity, diabetes, HTN, and hyperlipidemia in a general population. However, the relationship between serum magnesium and obesity in diabetic populations has been scarcely investigated and remains controversial. For example, similar to our findings, a study involving 119 adults with T2DM aged 25 to 65 years showed that the subjects with hypomagnesemia had significantly higher BMI than those with normal serum magnesium [33]. Babikr et al. [34] also reported a strongly negative correlation between serum magnesium and BMI in diabetic patients (r=–0.234, P<0.0001), which was also consistent with our findings.

However, in contrast to our findings and the studies mentioned above [33,34], Pokharel et al. [35] found no significant association between serum magnesium and the development of obesity in patients with T2DM. Additionally, in a study of 65 patients with newly diagnosed T2DM, no significant differences were observed in BMI, WC, and WHR values between those with and without magnesium deficiency [36]. Therefore, we conducted the current study in T2DM population to investigate the real relationship between serum magnesium and obesity as well as abdominal obesity. In fact, we observed a strong association of decreased serum magnesium with obesity and abdominal obesity in T2DM subjects. Moreover, the RCS plot also provided a significant non-linear negative correlation between serum magnesium and obesity as well as abdominal obesity, especially when serum magnesium levels were higher than 0.89 mmol/L.

Furthermore, in addition to BMI, our data showed that an increase in serum magnesium levels was also significantly associated with a decrease in WC, WHR, and WHtR values, which are well-known indicators of abdominal obesity. Similar to our results, Hirschler et al. [37] also found a significantly negative correlation between serum magnesium and WC, and a strong association of hypomagnesemia with central obesity in indigenous Argentinean children living at high altitude. Similar results were observed in another study including 428 adult patients with T2DM, in which a significant correlation between serum magnesium levels and WHR was found [38]. Since abdominal obesity is strongly linked to metabolic diseases [14], the significant correlation between serum magnesium levels and abdominal obesity suggests that serum magnesium may play an important role in metabolic regulation.

Interestingly, a remarkable decrease in the prevalence of severe obesity and abdominal obesity was found in the higher quintiles of serum magnesium, which suggested that low levels of serum magnesium may be a predictor of the progression to severe obesity and abdominal obesity in T2DM. Similar to our study, an investigation in Brazil found that serum magnesium levels significantly decreased with advancing obesity [39]. Moreover, Zaakouk et al. [40] also observed a significantly negative association between serum magnesium levels and the degree of obesity severity in children and adolescents. Furthermore, in a 30-year prospective cohort study, the most significantly negative correlation between magnesium intake and BMI was found in the highest quintile of magnesium intake [41]. Therefore, these studies including our study confirmed that high levels of serum magnesium have a notably ameliorative effect on the severity of obesity in different populations including T2DM population.

In addition, our study further revealed that serum CRP levels not only noticeably increased in T2DM subjects with obesity and abdominal obesity, but also obviously and positively associated with the severity of obesity and abdominal obesity. Furthermore, serum CRP in patients with T2DM was observed to be significantly and negatively correlated with serum magnesium. Additionally, 7.54% and 10.42% of the effect of serum magnesium on obesity and abdominal obesity was attributed to CRP-mediated effects. Thus, the protective role of serum magnesium in obesity is in part due to its anti-inflammatory effects. CRP, a well-known and reliable indicator of systemic inflammation, has been shown to remarkably correlate with obesity [42]. Furthermore, a previous study demonstrated that abdominal obesity exhibited the strongest remarkable correlation with CRP levels compared to other MetS components [43].

Obesity has been considered to be a disease characterized by chronic low-grade inflammation, and oxidative stress has been shown to be a crucial pathological mechanism in the development of obesity [44]. In a 20-year follow-up study, Kim et al. [45] observed a significantly negative correlation between serum magnesium and CRP levels in population without diabetes at baseline, which further corroborated our findings in patients with T2DM. Meanwhile, basic researches have also demonstrated the anti-oxidative stress and anti-inflammatory effects of magnesium by acting as a physiological inhibitor of xanthine-oxidase and nicotinamide adenine dinucleotide phosphate oxidase and as a cofactor of adenosine triphosphate to inhibit the production of reactive oxygen species [46,47]. Therefore, we speculated that serum magnesium protects against the onset and progression of obesity and abdominal obesity in T2DM patients, mainly through its anti-inflammatory role.

Moreover, a significant decline in FPG, 2hr PPG, TG, and HbA1c levels with increasing quintiles of serum magnesium was identified in the present study. Additionally, as serum magnesium quintiles increased, 2hr CP levels were elevated considerably, indicating that serum magnesium probably plays an important role in preserving insulin function. In fact, various studies have demonstrated that magnesium ion, serving as a cofactor for rate-limiting enzymes such as tyrosine kinase and glucokinase, plays a key role in the regulation of insulin secretion and function, and glycolysis [48]. Accordingly, it has been reported that FPG and PPG levels were significantly high in T2DM with hypomagnesemia [38]. In addition, a randomized controlled trial revealed that magnesium supplementation dramatically increased the levels of FCP and fasting insulin in overweight participants [49]. Therefore, high levels of serum magnesium may influence the regulation of energy metabolism, particularly glucose metabolism, to alleviate the development of obesity and abdominal obesity.

There were several limitations in the current study. Firstly, it was a single-center, cross-sectional study, thus the causal relationship between serum magnesium and obesity and abdominal obesity could not be established and needed to be further validated in future prospective studies. Additionally, dietary intake and exercise habits can influence obesity and abdominal obesity, whereas data on the specific diets and exercise patterns of our study participants were not available. In addition, we did not measure more direct parameters of oxidative stress to investigate their association with the occurrence and development of obesity and abdominal obesity. Finally, only hospitalized patients with T2DM were included in our study, which limited the generalizability of our findings to other populations such as diabetic patients from the community.

In conclusion, this real-world study with a large sample provided clear clinical evidence that serum magnesium not only significantly and negatively associated with obesity and abdominal obesity, but also linked with the severity of obesity and abdominal obesity in patients with T2DM, which probably attributed to its metabolic modulation, anti-oxidative stress, and anti-inflammatory effects. Serum magnesium may be used to assess the risk of obesity and abdominal obesity in T2DM subjects.

SUPPLEMENTARY MATERIALS

Supplementary materials related to this article can be found online at https://doi.org/10.4093/dmj.2023.0401.

Supplementary Fig. 1.

Comparisons of the prevalence of obesity and abdominal obesity stratified by sex, age, and duration of diabetes (DD). (A) Comparison of the prevalence of obesity and abdominal obesity between men and women after adjusting for age and DD (P=0.509 in obesity, P<0.001 in abdominal obesity). (B) Comparison of the prevalence of obesity and abdominal obesity stratified by age after adjusting for sex and DD (P<0.001 for trend in obesity, P=0.002 for trend in abdominal obesity). (C) Comparison of the prevalence of obesity and abdominal obesity stratified by DD after adjusting for age and sex (P=0.013 for trend in obesity, P=0.019 for trend in abdominal obesity).

dmj-2023-0401-Supplementary-Fig-1.pdf
Supplementary Fig. 2.

Linear relationship between serum magnesium levels and body mass index (BMI), waist circumference (WC), waist-to-hip ratio (WHR), and waist-to-height ratio (WHtR). (A) Correlation of serum magnesium levels with BMI after adjusting for gender, age, and duration of diabetes (DD) (R=–0.077, P<0.001). (B) Correlation of serum magnesium levels with WC after adjusting for gender, age, and DD (R=–0.074, P<0.001). (C) Correlation of serum magnesium levels with WHR after adjusting for gender, age, and DD (R=–0.089, P<0.001). (D) Correlation of serum magnesium levels with WHtR after adjusting for gender, age, and DD (R=–0.078, P<0.001).

dmj-2023-0401-Supplementary-Fig-2.pdf
Supplementary Fig. 3.

Association of serum magnesium with obesity and abdominal obesity analyzed by restrictive cubic spline (RCS). (A) Association between serum magnesium and obesity analyzed by RCS after adjusting for confounders in model 6. (B) Association between serum magnesium and abdominal obesity analyzed by RCS after adjusting for confounders in model 6. The solid line and the shaded portion represented the estimated odds ratio and its 95% confidence interval. Knots are at the 10th, 50th, and 90th centiles percentiles for serum magnesium.

dmj-2023-0401-Supplementary-Fig-3.pdf

Notes

CONFLICTS OF INTEREST

No potential conflict of interest relevant to this article was reported.

AUTHOR CONTRIBUTIONS

Conception or design: L.X.L.

Acquisition, analysis, or interpretation of data: M.R.X., A.P.W., Y.J.W., J.X.L.

Drafting the work or revising: L.X.L., M.R.X., Y.J.W., J.X.L.

Final approval of the manuscript: all authors.

FUNDING

T his work was supported by the National Natural Science Foundation of China (81170759, 81770813, and 82070866), the National Key Research and Development Plan (2018YFC1314900 and 2018YFC1314905), the Shanghai Research Center for Endocrine and Metabolic Diseases (2022ZZ01002), and Shanghai Municipal Key Clinical Specialty.

Acknowledgements

We thank the other investigators, the staff and all the patients of the present study for their invaluable contributions.

References

1. de Baaij JH, Hoenderop JG, Bindels RJ. Magnesium in man: implications for health and disease. Physiol Rev 2015;95:1–46.
2. Kostellow AB, Morrill GA. Iron-catalyzed lipid peroxidation in aortic cells in vitro: protective effect of extracellular magnesium. Atherosclerosis 2004;175:15–22.
3. Maier JA, Castiglioni S, Locatelli L, Zocchi M, Mazur A. Magnesium and inflammation: advances and perspectives. Semin Cell Dev Biol 2021;115:37–44.
4. Morais JB, Severo JS, Santos LR, de Sousa Melo SR, de Oliveira Santos R, de Oliveira AR, et al. Role of magnesium in oxidative stress in individuals with obesity. Biol Trace Elem Res 2017;176:20–6.
5. Leenders NH, Vermeulen EA, van Ballegooijen AJ, Hoekstra T, de Vries R, Beulens JW, et al. The association between circulating magnesium and clinically relevant outcomes in patients with chronic kidney disease: a systematic review and meta-analysis. Clin Nutr 2021;40:3133–47.
6. Lopez-Baltanas R, Rodriguez-Ortiz ME, Diaz-Tocados JM, Martinez-Moreno JM, Membrives C, Rodelo-Haad C, et al. Dietary Mg supplementation decreases oxidative stress, inflammation, and vascular dysfunction in an experimental model of metabolic syndrome with renal failure. Antioxidants (Basel) 2023;12:283.
7. Prabodh S, Prakash DS, Sudhakar G, Chowdary NV, Desai V, Shekhar R. Status of copper and magnesium levels in diabetic nephropathy cases: a case-control study from South India. Biol Trace Elem Res 2011;142:29–35.
8. Hamdan HZ, Nasser NM, Adam AM, Saleem MA, Elamin MI. Serum magnesium, iron and ferritin levels in patients with diabetic retinopathy attending Makkah Eye Complex, Khartoum, Sudan. Biol Trace Elem Res 2015;165:30–4.
9. Wang S, Hou X, Liu Y, Lu H, Wei L, Bao Y, et al. Serum electrolyte levels in relation to macrovascular complications in Chinese patients with diabetes mellitus. Cardiovasc Diabetol 2013;12:146.
10. Barbagallo M, Veronese N, Dominguez LJ. Magnesium in type 2 diabetes mellitus, obesity, and metabolic syndrome. Nutrients 2022;14:714.
11. Zhang H, Yan C, Yang Z, Zhang W, Niu Y, Li X, et al. Alterations of serum trace elements in patients with type 2 diabetes. J Trace Elem Med Biol 2017;40:91–6.
12. Haslam DW, James WP. Obesity. Lancet 2005;366:1197–209.
13. Chen MY, Zhao CC, Li TT, Zhu Y, Yu TP, Bao YQ, et al. Serum uric acid levels are associated with obesity but not cardio-cerebrovascular events in Chinese inpatients with type 2 diabetes. Sci Rep 2017;7:40009.
14. Alberti KG, Zimmet P, Shaw J, ; IDF Epidemiology Task Force Consensus Group. The metabolic syndrome: a new worldwide definition. Lancet 2005;366:1059–62.
15. Hassan SA, Ahmed I, Nasrullah A, Haq S, Ghazanfar H, Sheikh AB, et al. Comparison of serum magnesium levels in overweight and obese children and normal weight children. Cureus 2017;9e1607.
16. Bertinato J, Wu Xiao C, Ratnayake WM, Fernandez L, Lavergne C, Wood C, et al. Lower serum magnesium concentration is associated with diabetes, insulin resistance, and obesity in South Asian and white Canadian women but not men. Food Nutr Res 2015;59:25974.
17. Alasfar F, Ben-Nakhi M, Khoursheed M, Kehinde EO, Alsaleh M. Selenium is significantly depleted among morbidly obese female patients seeking bariatric surgery. Obes Surg 2011;21:1710–3.
18. Chu C, Zhao W, Zhang Y, Li L, Lu J, Jiang L, et al. Low serum magnesium levels are associated with impaired peripheral nerve function in type 2 diabetic patients. Sci Rep 2016;6:32623.
19. Xing B, Xu X, Li C, Zhao Y, Wang Y, Zhao W. Reduced serum magnesium levels are associated with the occurrence of retinopathy in patients with type 2 diabetes mellitus: a retrospective study. Biol Trace Elem Res 2022;200:2025–32.
20. Huang JH, Lu YF, Cheng FC, Lee JN, Tsai LC. Correlation of magnesium intake with metabolic parameters, depression and physical activity in elderly type 2 diabetes patients: a cross-sectional study. Nutr J 2012;11:41.
21. Zhang ZH, Ke JF, Lu JX, Liu Y, Wang AP, Li LX. Serum retinolbinding protein levels are associated with nonalcoholic fatty liver disease in Chinese patients with type 2 diabetes mellitus: a real-world study. Diabetes Metab J 2022;46:129–39.
22. Ke JF, Wang JW, Lu JX, Zhang ZH, Liu Y, Li LX. Waist-to-height ratio has a stronger association with cardiovascular risks than waist circumference, waist-hip ratio and body mass index in type 2 diabetes. Diabetes Res Clin Pract 2022;183:109151.
23. Ma YL, Jin CH, Zhao CC, Ke JF, Wang JW, Wang YJ, et al. Waist-to-height ratio is a simple and practical alternative to waist circumference to diagnose metabolic syndrome in type 2 diabetes. Front Nutr 2022;9:986090.
24. Xu MR, Jin CH, Lu JX, Li MF, Li LX. High-normal unconjugated bilirubin is associated with decreased risk of chronic kidney disease in type 2 diabetes: a real-world study. Diabetes Metab Res Rev 2023;39e3672.
25. Ke JF, Wang JW, Zhang ZH, Chen MY, Lu JX, Li LX. Insulin therapy is associated with an increased risk of carotid plaque in type 2 diabetes: a real-world study. Front Cardiovasc Med 2021;8:599545.
26. Li MF, Ren Y, Zhao CC, Zhang R, Li LX, Liu F, et al. Prevalence and clinical characteristics of lower limb atherosclerotic lesions in newly diagnosed patients with ketosis-onset diabetes: a cross-sectional study. Diabetol Metab Syndr 2014;6:71.
27. Liu F, Chen S, Zhao W, Chen M, Ke J, Zhang Z, et al. Urine uric acid excretion levels are positively associated with obesity and abdominal obesity in type 2 diabetes patients without chronic kidney disease. Diabetes Metab Syndr Obes 2021;14:4691–703.
28. Kushner I, Antonelli MJ. What should we regard as an “elevated” C-reactive protein level? Ann Intern Med 2015;163:326.
29. Muneyyirci-Delale O, Nacharaju VL, Dalloul M, Altura BM, Altura BT. Serum ionized magnesium and calcium in women after menopause: inverse relation of estrogen with ionized magnesium. Fertil Steril 1999;71:869–72.
30. Masood N, Baloch GH, Ghori RA, Memon IA, Memon MA, Memon MS. Serum zinc and magnesium in type-2 diabetic patients. J Coll Physicians Surg Pak 2009;19:483–6.
31. Shi Z, Abou-Samra AB. Association of low serum magnesium with diabetes and hypertension: findings from Qatar Biobank study. Diabetes Res Clin Pract 2019;158:107903.
32. Li Q, Chen Q, Zhang H, Xu Z, Wang X, Pang J, et al. Associations of serum magnesium levels and calcium-magnesium ratios with mortality in patients with coronary artery disease. Diabetes Metab 2020;46:384–91.
33. Ozcaliskan Ilkay H, Sahin H, Tanriverdi F, Samur G. Association between magnesium status, dietary magnesium intake, and metabolic control in patients with type 2 diabetes mellitus. J Am Coll Nutr 2019;38:31–9.
34. Babikr WG, Shalayel MH, Abdelraheem AH, Elhussein AB. Impact of body mass index on hypomagnesemia and hypocalcemia in type 2 diabetic patients. Biomed Pharm J 2016;9:933–40.
35. Pokharel DR, Khadka D, Sigdel M, Yadav NK, Kafle R, Sapkota RM, et al. Association of serum magnesium level with poor glycemic control and renal functions in Nepalese patients with type 2 diabetes mellitus. Diabetes Metab Syndr 2017;11 Suppl 1:S417–23.
36. Ruhul-Kabir M, Kamrul-Hasan AB, Burhan-Uddin SA, Majumder M, Rahman H, Nazmul-Islam AF. Serum magnesium status and its correlation with insulin resistance in newly diagnosed patients with type 2 diabetes mellitus. Sri Lanka J Diab Endocrinol Metabol 2019;9:11–7.
37. Hirschler V, Esteban M, Gonzalez C, Molinari C, Castano L. Association between waist circumference and magnesium and uric acid in indigenous Argentinean children living at high altitude. Cardiovasc Hematol Agents Med Chem 2016;14:134–40.
38. Wolide AD, Zawdie B, Alemayehu T, Tadesse S. Evaluation of serum ferritin and some metal elements in type 2 diabetes mellitus patients: comparative cross-sectional study. Diabetes Metab Syndr Obes 2016;9:417–24.
39. Johansson HE, Zethelius B, Ohrvall M, Sundbom M, Haenni A. Serum magnesium status after gastric bypass surgery in obesity. Obes Surg 2009;19:1250–5.
40. Zaakouk AM, Hassan MA, Tolba OA. Serum magnesium status among obese children and adolescents. Egypt Pediatr Assoc Gaz 2016;64:32–7.
41. Lu L, Chen C, Yang K, Zhu J, Xun P, Shikany JM, et al. Magnesium intake is inversely associated with risk of obesity in a 30- year prospective follow-up study among American young adults. Eur J Nutr 2020;59:3745–53.
42. Choi J, Joseph L, Pilote L. Obesity and C-reactive protein in various populations: a systematic review and meta-analysis. Obes Rev 2013;14:232–44.
43. Chapidze G, Dolidze N, Enquobahrie DA, Kapanadze S, Latsabidze N, Williams MA. Metabolic syndrome and C-reactive protein among cardiology patients. Arch Med Res 2007;38:783–8.
44. Saltiel AR, Olefsky JM. Inflammatory mechanisms linking obesity and metabolic disease. J Clin Invest 2017;127:1–4.
45. Kim DJ, Xun P, Liu K, Loria C, Yokota K, Jacobs DR Jr, et al. Magnesium intake in relation to systemic inflammation, insulin resistance, and the incidence of diabetes. Diabetes Care 2010;33:2604–10.
46. Barbagallo M, Dominguez LJ. Magnesium and aging. Curr Pharm Des 2010;16:832–9.
47. Pasternak K, Kocot J, Horecka A. Biochemistry of magnesium. J Elementol 2010;15:601–16.
48. Kostov K. Effects of magnesium deficiency on mechanisms of insulin resistance in type 2 diabetes: focusing on the processes of insulin secretion and signaling. Int J Mol Sci 2019;20:1351.
49. Chutia H, Lynrah KG. Association of serum magnesium deficiency with insulin resistance in type 2 diabetes mellitus. J Lab Physicians 2015;7:75–8.

Article information Continued

Fig. 1.

Comparisons of serum magnesium levels in different groups. (A) Comparison of serum magnesium levels between men and women with type 2 diabetes mellitus after adjusting for age and duration of diabetes (DD) (P<0.001). (B) Comparison of serum magnesium levels stratified by the degree of obesity after adjusting for age, sex, and DD (P<0.001 for trend). (C) Comparison of serum magnesium levels stratified by the degree of abdominal obesity after adjusting for age, sex, and DD (P<0.001 for trend). (D) Comparison of serum magnesium levels stratified by age after adjusting for DD and sex (P<0.001 for trend). (E) Comparison of serum magnesium levels stratified by DD after adjusting for age and sex (P<0.001 for trend).

Fig. 2.

Comparisons of the prevalence of obesity and abdominal obesity among serum magnesium quintiles. (A) Comparison of the prevalence of obesity and abdominal obesity stratified by serum magnesium quintiles after adjusting for age, sex, and duration of diabetes (DD) (P<0.001 for trend in obesity, P=0.001 for trend in abdominal obesity). (B) Comparison of the prevalence of obesity and abdominal obesity in low, normal, high serum magnesium group after adjusting for age, sex, and DD (P=0.030 for trend in obesity, P=0.088 for trend in abdominal obesity). (C) Comparison of the prevalence of mild, moderate, and severe obesity across serum magnesium quintiles after adjusting for age, sex, and DD (P<0.001 for trend). (D) Comparison of the prevalence of mild, moderate, and severe abdominal obesity across serum magnesium quintiles after adjusting for age, sex, and DD (P<0.001 for trend).

Fig. 3.

Role of C-reactive protein (CRP) in the relationship between serum magnesium and obesity/abdominal obesity. (A) Comparison of serum CRP levels between the patients with and without obesity/abdominal obesity after controlling for gender, age, and duration of diabetes (DD) (P<0.001 for trend in obesity and abdominal obesity). (B) Comparison of serum CRP levels in mild, moderate, and severe obesity/abdominal obesity after controlling for gender, age, and DD (P=0.025 for trend in obesity, P<0.001 for trend in abdominal obesity). (C) Comparison of serum CRP levels across serum magnesium quintiles after adjusting for gender, age, and DD (P=0.046 for trend). (D) The mediation effects of CRP on the association between serum magnesium and obesity (P=0.016). (E) The mediation effects of CRP on the association between serum magnesium and abdominal obesity (P=0.004). ACME, average causal mediation effect; ADE, average direct effects.

Table 1.

Characteristics of the subjects according to serum magnesium quintiles

Variable Q1 (n=1,595) Q2 (n=1,899) Q3 (n=1,489) Q4 (n=1,532) Q5 (n=1,495) P valuea P valueb
Serum magnesium, mmol/L <0.82 0.82–0.87 0.88–0.91 0.92–0.96 >0.97 - -
Male sex 782 (49.0) 1,046 (55.1) 858 (57.6) 873 (57.0) 848 (56.7) <0.001 <0.001
Age, yr 58±13 59±13 59±12 61±12 62±12 <0.001 <0.001
DD, moc 108 (48–156) 96 (36–156) 96 (36–156) 96 (36–156) 84 (36–144) 0.003 0.011
Smoking 458 (28.7) 570 (30.0) 468 (31.4) 467 (30.5) 436 (29.2) 0.498 0.468
Drinking 253 (15.9) 305 (16.1) 276 (18.5) 250 (16.3) 256 (17.1) 0.257 0.451
HTN 889 (55.7) 1,014 (53.4) 828 (55.6) 812 (53.0) 833 (55.7) 0.300 0.001
IIAs 1,226 (76.9) 1,428 (75.2) 1,032 (69.3) 969 (63.3) 896 (59.9) <0.001 <0.001
SUs 1,026 (64.3) 1,245 (65.6) 962 (64.6) 1,017 (66.4) 989 (66.2) 0.685 0.954
Metformin 1,058 (66.3) 1,170 (61.6) 900 (60.4) 849 (55.4) 774 (51.8) <0.001 <0.001
SBP, mm Hg 133±17 132±17 132±17 132±17 132±18 0.759 0.051
DBP, mm Hg 81±9 80±9 80±9 79±9 80±9 0.006 0.850
BMI, kg/m2 25.37±3.84 25.19±3.46 25.07±3.33 24.81±3.38 24.67±3.28 0.752 0.871
WC, cm 90.41±10.95 90.51±10.40 89.84±10.05 89.21±10.07 88.97±10.33 <0.001 0.630
WHtR 0.55±0.07 0.55±0.06 0.54±0.06 0.54±0.06 0.54±0.06 <0.001 0.370
WHR 0.92±0.07 0.92±0.07 0.92±0.06 0.92±0.06 0.91±0.06 <0.001 0.820
FPG, mmol/Lc 8.4 (6.7–10.7) 8.0 (6.4–10.1) 7.7 (6.2–9.5) 7.3 (5.9–9.2) 7.0 (5.8–8.5) <0.001 <0.001
2hr PPG, mmol/Lc 13.9 (11.1–17.3) 13.9 (10.7–17.2) 13.2 (9.9–16.6) 12.4 (9.5–15.9) 12.1 (9.3–15.3) <0.001 <0.001
HbA1c, % 9.50±2.29 9.14±2.18 8.79±2.07 8.38±2.01 8.12±2.04 <0.001 <0.001
FCP, ng/mLc 1.81 (1.09–2.56) 1.71 (1.11–2.49) 1.82 (1.15–2.57) 1.83 (1.24–2.61) 1.95 (1.24–2.76) <0.001 0.087
2hr CP, ng/mLc 3.45 (2.05–5.58) 3.85 (2.21–5.80) 4.08 (2.41–6.03) 4.44 (2.75–6.25) 4.56 (2.78–6.42) <0.001 <0.001
HOMA2-IR 1.59 (0.95–2.30) 1.49 (0.96–2.17) 1.53 (0.95–2.22) 1.57 (1.00–2.20) 1.61 (1.03–2.27) 0.078 0.807
TG, mmol/Lc 1.52 (1.04–2.33) 1.44 (1.00–2.15) 1.39 (0.99–2.14) 1.39 (0.98–2.02) 1.38 (0.99–1.97) <0.001 0.699
TC, mmol/L 4.78±1.20 4.75±1.22 4.74±1.11 4.72±1.03 4.73±1.14 0.648 0.400
HDL-C, mmol/L 1.10±0.31 1.12±0.32 1.13±0.30 1.14±0.30 1.15±0.33 <0.001 0.901
LDL-C, mmol/L 2.96±0.95 3.01±0.94 3.05±0.94 3.05±0.89 3.07±0.97 0.007 0.502
ALT, U/Lc 17 (12–26) 18 (13–28) 19 (13–29) 19 (14–30) 21 (15–33) 0.534 0.252
Cr, μmol/Lc 62 (50–75) 65 (54–78) 66 (55–78) 68 (58–81) 71 (60–85) <0.001 0.008
SUA, μmol/Lc 317 (260–384) 314 (261–374) 310 (258–372) 311 (263–372) 319 (266–377) 0.067 0.953
UAE, mg/dayc 15 (8–53) 12 (7–37) 11 (7–28) 11 (6–26) 10 (6–26) <0.001 0.003
eGFR, mL/min/1.73 m2c 121 (96–146) 123 (93–136) 110 (92–131) 106 (88–126) 101 (83–120) <0.001 <0.001

Values are presented as number (%), mean±standard deviation, or median (interquartile range).

DD, duration of diabetes; HTN, hypertension; IIA, insulin or insulin analogue; SU, sulfonylurea; SBP, systolic blood pressure; DBP, diastolic blood pressure; BMI, body mass index; WC, waist circumference; WHtR, waist-to-height ratio; WHR, waist-to-hip ratio; FPG, fasting plasma glucose; 2hr PPG, 2-hour postprandial plasma glucose; HbA1c, glycosylated hemoglobin; FCP, fasting C-peptide; 2hr CP, 2-hour postprandial C-peptide; HOMA2-IR, updated homoeostasis model assessment of insulin sensitivity/insulin resistance; TG, total triglyceride; TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; ALT, alanine aminotransferase; Cr, creatinine; SUA, serum uric acid; UAE, urinary albumin excretion; eGFR, estimated glomerular filtration rate.

a

The P values were not adjusted for sex for the trend,

b

The P values were adjusted for sex for the trend,

c

The Mann-Whitney U test was applied.

Table 2.

The association of serum magnesium levels with obesity and abdominal obesity

Variable OR 95% CI P value
Obesity
 Model 1 0.879 0.841–0.919 <0.001
 Model 2 0.888 0.849–0.928 <0.001
 Model 3 0.899 0.859–0.940 <0.001
 Model 4 0.919 0.877–0.963 <0.001
 Model 5 0.922 0.881–0.966 0.001
 Model 6 0.885 0.835–0.938 <0.001
Abdominal obesity
 Model 1 0.881 0.840–0.923 <0.001
 Model 2 0.886 0.843–0.932 <0.001
 Model 3 0.895 0.851–0.942 <0.001
 Model 4 0.925 0.878–0.975 0.004
 Model 5 0.927 0.880–0.977 0.004
 Model 6 0.910 0.852–0.971 0.004

Model 1: unadjusted; Model 2: adjusted for age, gender, and duration of diabetes; Model 3: additionally adjusted for hypertension, smoking, and alcohol drinking; Model 4: additionally adjusted for the use of insulin or insulin analogues, metformin, and sulfonylurea; Model 5: additionally adjusted for systolic blood pressure and diastolic blood pressure; Model 6: further adjusted for alanine aminotransferase, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, total triglyceride, estimated glomerular filtration rate, creatinine, serum uric acid, urinary albumin excretion, fasting plasma glucose, 2-hour postprandial plasma glucose, glycosylated hemoglobin, fasting C-peptide, 2-hour postprandial C-peptide, and C-reactive protein.

OR, odds ratio; CI, confidence interval.

Table 3.

Association of serum magnesium quintiles with obesity and abdominal obesity

Variable OR (95% CI)
P values for trend
Q1 Q2 Q3 Q4 Q5
Obesity
 Model 1 1 0.979 (0.857–1.119) 0.909 (0.789–1.047) 0.788 (0.684–0.906) 0.740 (0.642–0.852) <0.001
 Model 2 1 0.990 (0.866–1.131) 0.921 (0.799–1.061) 0.805 (0.699–0.927) 0.761 (0.660–0.877) <0.001
 Model 3 1 1.023 (0.893–1.171) 0.937 (0.812–1.082) 0.843 (0.730–0.973) 0.786 (0.680–0.908) 0.001
 Model 4 1 1.049 (0.914–1.204) 0.962 (0.831–1.113) 0.889 (0.768–1.029) 0.844 (0.728–0.978) 0.020
 Model 5 1 1.057 (0.920–1.214) 0.963 (0.831–1.115) 0.887 (0.766–1.028) 0.854 (0.736–0.990) 0.023
 Model 6 1 1.093 (0.923–1.294) 0.947 (0.791–1.134) 0.870 (0.727–1.042) 0.768 (0.639–0.923) 0.001
Abdominal obesity
 Model 1 1 0.952 (0.822–1.102) 0.855 (0.733–0.997) 0.790 (0.678–0.919) 0.727 (0.625–0.847) <0.001
 Model 2 1 1.007 (0.865–1.173) 0.921 (0.785–1.081) 0.832 (0.710–0.975) 0.753 (0.643–0.882) 0.001
 Model 3 1 1.043 (0.894–1.217) 0.932 (0.792–1.095) 0.872 (0.743–1.024) 0.774 (0.660–0.908) 0.001
 Model 4 1 1.077 (0.921–1.261) 0.968 (0.821–1.142) 0.938 (0.796–1.106) 0,850 (0.720–1.002) 0.053
 Model 5 1 1.082 (0.925–1.267) 0.970 (0.822–1.144) 0.934 (0.793–1.101) 0.859 (0.728–1.014) 0.063
 Model 6 1 1.125 (0.925–1.367) 0.961 (0.784–1.178) 1.000 (0.817–1.225) 0.823 (0.671–1.010) 0.031

Model 1: unadjusted; Model 2: adjusted for age, gender, and duration of diabetes; Model 3: additionally adjusted for hypertension, smoking and alcohol drinking; Model 4: additionally adjusted for the use of insulin or insulin analogues, metformin, and sulfonylurea; Model 5: additionally adjusted for systolic blood pressure and diastolic blood pressure; Model 6: further adjusted for alanine aminotransferase, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglyceride, estimated glomerular filtration rate, creatinine, serum uric acid, urinary albumin excretion, fasting plasma glucose, 2-hour postprandial plasma glucose, glycosylated hemoglobin, fasting C-peptide, 2-hour postprandial C-peptide, and C-reactive protein.

OR, odds ratio; CI, confidence interval.