Clinical Nutrition 30 (2011) 359e364

Contents lists available at ScienceDirect

Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu

Original Article

Influence of magnesium status and magnesium intake on the blood glucose

control in patients with type 2 diabetesq
Cristiane Hermes Sales a, Lucia Fátima Campos Pedrosa b, Josivan Gomes Lima c,
Telma Maria Araújo Moura Lemos d, Célia Colli a, *
a
Department of Food and Experimental Nutrition, Faculty of Pharmaceutical Sciences, University of São Paulo, Av. Prof. Lineu Prestes, 580, Bl. 14. Butantã, São Paulo,
SP CEP 05508-000, Brazil
b
Department of Nutrition, Federal University of Rio Grande do Norte, Av. General Cordeiro de Farias, s/n, Petrópolis, Natal, RN 59010-180, Brazil
c
Department of Clinical Medicine, Federal University of Rio Grande do Norte, Av. General Cordeiro de Farias, s/n, Petrópolis, Natal, RN 59010-180, Brazil
d
Department of Clinical and Toxicological Analyses, Federal University of Rio Grande do Norte, Av. General Cordeiro de Farias, s/n, Petrópolis, Natal, RN 59010-180, Brazil

a r t i c l e i n f o s u m m a r y

Article history: Background & aims: This study was undertaken to assess magnesium intake and magnesium status in
Received 21 August 2010 patients with type 2 diabetes, and to identify the parameters that best predict alterations in fasting
Accepted 23 December 2010 glucose and plasma magnesium.
Methods: A cross-sectional study was carried out in patients with type 2 diabetes (n ¼ 51; 53.6  10.5 y)
Keywords: selected within the inclusion factors, at the University Hospital Onofre Lopes. Magnesium intake was
Magnesium
assessed by three 24-h recalls. Urine, plasma and erythrocytes magnesium, fasting and 2-h postprandial
Diabetes
glucose, HbA1, microalbuminuria, proteinuria, and serum and urine creatinine were measured.
Assessment
Glucose control
Results: Mean magnesium intake (9.37  1.76 mmol/d), urine magnesium (2.80  1.51 mmol/d), plasma
Kidney function magnesium (0.71  0.08 mmol/L) and erythrocyte magnesium (1.92  0.23 mmol/L) levels were low.
Discriminant analysis Seventy-seven percent of participants presented one or more magnesium status parameters below the
cut-off points of 3.00 mmol/L for urine, 0.75 mmol/L for plasma and 1.65 mmol/L for erythrocytes.
Subjects presented poor blood glucose control with fasting glucose of 8.1  3.7 mmol/L, 2-h postprandial
glucose of 11.1  5.1 mmol/L, and HbA1 of 11.4  3.0%. The parameters that influenced fasting glucose
were urine, plasma and dietary magnesium, while plasma magnesium was influenced by creatinine
clearance.
Conclusions: Magnesium status was influenced by kidney depuration and was altered in patients with
type 2 diabetes, and magnesium showed to play an important role in blood glucose control.
Ó 2011 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

1. Introduction

Magnesium is an essential mineral for the human body, prin-

Abbreviations: HUOL, University Hospital Onofre Lopes; BMI, body mass index;
WC, waist circumference; CCr, creatinine clearance; HbA1, glycated hemoglobin;
BDS, Brazilian Diabetes Society; D, difference between the observed average intake
by each subject; SDD, standard deviation of D; EAR, estimated average requirement.
q Conference presentation: Part of this study was presented orally at the 10th
National Congress of the Brazilian Food and Nutrition Society (São Paulo, Brazil,
September 2009), at the XVII Congress of the Brazilian Diabetes Society (Fortaleza,
Brazil, November 2009) and at the XV Congress of the Latin American Nutrition
Society (Santiago, Chile, November 2009), and their abstracts were published in the
Book of Abstracts.
* Corresponding author. Tel.: þ55 11 3091 3651; fax: þ55 11 3815 4410.
E-mail addresses: cristianehermes@yahoo.com.br (C.H. Sales), lpedrosa@ufrnet.
br (L.F.C. Pedrosa), josivanlima@gmail.com (J.G. Lima), telmaml@yahoo.com.br
(T.M.A.M. Lemos), cecolli@usp.br (C. Colli).
cipally because of its role in the regulation of cellular processes and
its function as a cofactor in a wide range of metabolic reactions.
Many enzymes that catalyze phosphorylation and dephosphory-
lation reactions, including those involved in glycolysis, are activated
by the formation of MgATPþ2 complexes, which are the real
substrates for these enzymes.1,2
Alterations in the distribution of magnesium within the body
have been associated with several diseases and especially diabetes,
a disorder representing a global public health problem of increas-
ingly serious concern.3,4 Although some epidemiological studies
have suggested that adequate magnesium intake reduces the risk of
development of type 2 diabetes, there are still contradictions with
respect to the role of low magnesium intake as a predictor factor for

0261-5614/$ e see front matter Ó 2011 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
doi:10.1016/j.clnu.2010.12.011
360 C.H. Sales et al. / Clinical Nutrition 30 (2011) 359e364
this disease.5e10 The importance of magnesium for individuals with
diabetes can be explained on the basis of maintenance of glucose
homeostasis along with activation of the factors involved in
sensitivity of tissues to insulin, the receptors of which are phos-
phorylated only in the presence of MgATPþ2.2,11
Some studies have shown that the magnesium intake by
patients with diabetes is often below recommended levels.12,13
Additionally, there is evidence that the magnesium status of
patients with diabetes tends to alter, and that low body concen-
trations of this mineral may influence the evolution of the disease
and generate further complications.14e17
Despite reports describing the occurrence of hypomagnesemia
among patients with diabetes,18,19 few investigations have consid-
ered dietary intake of magnesium in the Brazilian population, and
none has examined the levels of magnesium in patients with type 2
diabetes. Hence, the aim of the present study was to evaluate the
intake of magnesium and the levels of the mineral in urine, plasma
and erythrocytes in subjects with type 2 diabetes. Additionally,
attempts were made to identify those parameters that best predict
alterations in fasting glucose and plasma magnesium.
2. Materials and methods
The study was approved by the Ethics Committees on Research
of the University Hospital Onofre Lopes (HUOL, Natal, RN, Brazil)
and of the Faculty of Pharmaceutical Sciences, University of São
Paulo (protocol CAAE # 0005.0.294.018-06). Written informed
consent was obtained from all participants prior to the comm-
encement of the study.
2.1. Subjects
A cross-sectional study was carried out in patients with type 2
diabetes selected from those attending the Endocrinology Clinic at
HUOL as outpatients between February and June 2008. The size of
the study population was defined by the requirement to detect
a difference of 0.075 mmol/L in plasma magnesium with a power of
0.9070% and a 5% a-level, and was calculated using the Student
t-test for independent samples assuming a normal distribution of
plasma magnesium.
Fifty-one patients were selected, consecutively, on the basis
of medical records considering the following inclusion criteria:
   
1:73  24  h urine creatineðmg=dLÞ  24  h urine volumeðmL=sÞ
CCr ¼
serum creatineðmg=dLÞ
(a) medical diagnosis of type 2 diabetes; (b) age range 25e65 years;
(c) non-pregnant/non-lactating women; (d) absence of kidney
failure as defined by the levels of serum creatinine (<124 mmol/L for
women; <133 mmol/L for men); (e) absence of digestive, thyroid,
congenital and infectious diseases; (f) no recent history of alcohol
abuse, use of vitamin and mineral supplements or medication that
could interfere in the analysis of magnesium (except for antidia-
betic and antihypertensive drugs).
2.2. Study design
The selected participants were requested to visit the hospital
three times during the period of a month and were submitted on
each occasion to a 24-h food recall. Patients were formally invited
to participate in the study during their first hospital visit, following
which their arterial blood pressure was measured by a doctor and
anthropometric measurements were assessed by a responsible
researcher. Subjects were also submitted to an interview, which
was based on a previously structured questionnaire, applied by
a researcher. At the second visit, patients delivered urine samples
that they had collected (in mineral-free flasks) over the previous
24-h, and blood samples were collected by vein puncture from the
12e14 h fasting subjects. In order to determine postprandial serum
glucose, a further blood sample was collected 2-h after the break-
fast. During the third visit, patients received personal nutritional
orientation together with the results of the biochemical analyses
that would serve as references for subsequent medical follow-up.
2.3. Blood pressure and anthropometric measurements
All measurements were made in duplicate. Blood pressure was
determined using a mercury sphygmomanometer and readings
were taken from the upper left arm with the patient sitting down.
Body weight was assessed using Plenna (Sao Paulo, SP, Brazil)
calibrated digital scales (0.1 kg precision) with patients wearing
light clothes and no shoes. Height was evaluated (0.1 mm precision)
using a Cardiomed (Curitiba, PR, Brazil) stadiometer. Body mass
index (BMI) was expressed as the quotient between weight (kg)
and height squared (m2). Waist circumference (WC) was estimated
at the end of a normal expiration using a Cardiomed non-extend-
ible tape held in a horizontal plane around the abdomen at the level
of the iliac crest. Patients were classified according to BMI and WC
following the recommendations of the World Health Organization
as adopted by the American Diabetes Association.3
2.4. Biochemical analyses
All biochemical analyses were carried out in triplicate with a 10%
variation limit set as the criterion for repetition of the assay. Kidney
function was determined through the analysis of urine albumin,
urine protein, and urine and serum creatinine. Microalbuminuria
was measured turbidimetrically using a Biosystems (Barcelona,
Spain) kit, while proteinuria was assessed from the end point of the
pyrogallol red-molybdenum(VI) complex reaction. The levels of
creatinine in urine (24-h collection) and serum were estimated
using the alkaline picrate method without precipitation (Jaffe’s
reaction) with the aid of Labtest Diagnostica (Lagoa Santa, MG,
Brazil) kits. Creatinine clearance (CCr; mL/s/1.73 m2) was calculated
from the expression:
body surface area m2 (1)
Serum fasting glucose and 2-h postprandial glucose were
measured using the glucose oxidase enzyme method, while glycated
hemoglobin (HbA1) was determined by ion-exchange chromatog-
raphy, both analyses being performed with the aid of Labtest Diag-
nostica kits. Glycemic control was assessed on the basis of the
glycemia standards proposed by the Brazilian Diabetes Society (BDS)20
for the treatment of patients with type 2 diabetes, namely: fasting
glucose <6.11 mmol/L, 2-h postprandial glucose <7.77 mmol/L, and
HbA1 < 9.0%.
Magnesium status was evaluated directly by measuring the levels
of the mineral in urine, plasma and erythrocytes by flame atomic
absorption spectrometry (AAnalyst 100; Perkin Elmer, Norwalk, CT,
USA) according to previously standardized and validated protocols.21
Flasks and glassware were demineralized prior to analyses, and the
precision and accuracy of the methods were verified using certified
standards (Trace Elements Serum L-I and Urine Blank; Seronorm,
Billingstad, Norway) with urine, plasma and erythrocyte pools being
 C.H. Sales et al. / Clinical Nutrition 30 (2011) 359e364
employed as secondary standards. The magnesium concentration
cut-off points were: (a) urine 3.00e5.00 mmol/d,22 (b) plasma
0.75e1.05 mmol/L,1 and (c) erythrocytes 1.65e2.65 mmol/L.22
2.5. Magnesium intake
The 24-h food recalls were conducted by trained nutritionists
who employed an album containing images of food and utensils to
guide patients in evaluating the proportions of food consumed at
each meal. Data were analyzed using the NutriQuanti On-line
Computerized System (http://www.nutriquanti.com.br) in order to
estimate magnesium intake. The methods and recommendations
described in the Dietary Reference Intakes of the United States
Institute of Medicine were employed in estimating the magnesium
intake.23,24 The probability of adequate magnesium intake was
calculated from the ratio between D and SDD, which D is the
difference between the observed average intake by each subject and
the Estimated Average Requirement (EAR) taking into account the
life stage and gender group of the individual, and SDD is the standard
deviation of D calculated by taking into account the standard devi-
ation of the distribution of intake of the reference group and the
standard deviation of the data obtained from the three food recalls.
2.6. Statistical analyses
Statistical analyses were performed using SPSS software (Chi-
cago, IL, USA) version 15.0. The normality of data was tested using
the KolmogoroveSmirnov test. For normally distributed data, the
Student t-test and the Pearson correlation were employed, respec-
tively, to compare mean values and to evaluate the association
between parameters. For data that were not normally distributed,
mean values were compared using ManneWhitney’s test. For these
test were used the level of significance a was established at 5%.
In order to reduce the errors associated with dietary measure-
ments, the values were adjusted according to the total energy
intake using the residual method,25 and according to the intra-
individual variation calculated from the mean value of the
components of the analyses of variance.26
Stepwise discriminant analysis (Wilk’s lambda method) was
employed to identify the best parameters for predicting increased
fasting glucose and plasma magnesium deficiency. The parameters
included in the analysis were (a) gender, microalbuminuria,
proteinuria, CCr, magnesium concentration in urine, plasma and
erythrocytes, magnesium intake, carbohydrate, fat, and protein;
and (b) gender, microalbuminuria, proteinuria, CCr, fasting glucose,
2-h postprandial glucose, HbA1, magnesium intake, carbohydrate,
fat, and protein. Parameters that were not normally distributed
were submitted to natural log transformation. Box’s M test was
employed to verify the homogeneity of covariances. Assuming that
the variances were equivalent, the predicting parameters were
filtered and the discriminant functions were defined by stepwise
discriminant analysis with a set at 10%.
Patients were stratified according to fasting glucose (high
6.11 mmol/L; normal <6.11 mmol/L) and plasma magnesium
(deficient <0.75 mmol/L; adequate 0.75 mmol/L) levels. The cut-
off points were defined by goal suggested by BDS for the treatment
of patients with type 2 diabetes,20 and by the lower plasma Mg
level which the patient commonly do not have signs and symptoms
of Mg deficiency.1
3. Results
The baseline characteristics of the study population (n ¼ 51) are
shown in Table 1. The main drugs prescribed were oral antidiabetic
biguanides alone (metformin; 66.7%), followed by sulfonylureas
361
alone (31.4%) and a metformineglibenclamide combination (3.9%),
while the antihypertensive drugs used were hydrochlorothiazide
(23.5%) and captopril (45.1%). Half of the patients (51.0%) received
insulin treatment, often accompanied by antihyperglycemic drugs
such as metformin (29.4%), sulfonylurea (3.9%) and others (3.9%).
These drugs were prescribed because alone these are available in the
Brazilian Public Health System. With respect to nutritional status,
most of the subjects exhibited high BMI and high WC (Table 1).
Twelve of the 51 subjects presented urine plasma albumin levels
30 mg/d (mean 160.4  88.2 mg/d), possibly indicating initial
stages of kidney alteration (Table 2). Glycemic control was generally
unsatisfactory according to BDS standards for patients with type 2
diabetes20 in that a large number of subjects presented high levels of
fasting glucose (55%), postprandial glucose (61%) and HbA1 (80%).
Regarding urine, plasma and erythrocyte magnesium status,
77% of subjects (9 men, 30 women) presented values for one or
more of the parameters that were below reference levels. A large
number of patients (47%) exhibited low levels of urine and plasma
magnesium, while erythrocyte magnesium was normal. Low
concentrations of magnesium in urine, plasma and erythrocytes
were detected in 12% of subjects.
In general, magnesium intake by the study population was low
(Table 2). On the basis of individual assessment, none of the
subjects showed a probability of adequate magnesium intake
greater than 85%. On the other hand, 22 subjects showed an intake
of the mineral that was lower than the 15th percentile (Fig. 1).
When the subjects were stratified according to fasting glucose
levels (<6.11 or 6.11 mmol/L), significant differences were
observed with respect to plasma and erythrocyte magnesium.
Additionally, when the subjects were stratified according to plasma
magnesium levels (<0.75 or 0.75 mmol/L), significant differences
were observed regarding HbA1 (Table 2). There was a clear asso-
ciation between magnesium levels and glycemic control as shown
by the correlation coefficients between plasma magnesium vs.
fasting glucose (r ¼ 0.28, p ¼ 0.046), plasma magnesium vs. 2-h
postprandial glucose (r ¼ 0.32; p ¼ 0.021) and urine magnesium
vs. fasting glucose (r ¼ 0.291; p ¼ 0.038).
It is important to emphasize that there were no significant
differences between non-medicated patients and those receiving
insulin, metformin and diuretic drugs. Moreover, there were no
significant differences between individuals with and without
microalbuminuria (data not shown). In contrast, there were
significant differences between patients with hypomagnesemia
and normal subjects with respect to CCr (Table 2).
The results of discriminant analysis revealed that the most
important indicators for the classification of normal and hypergly-
cemic subjects were urine, plasma and dietetic magnesium, while
CCr was the only variable that influenced plasma magnesium levels.
Such relationships could be expressed by the polynomial equations;
Yfasting glucose ¼ 11.865 þ [14.425  plasma magnesium (mmol/
L)]  [0.479  urine magnesium (mmol/d)] þ [0.306  magnesium
intake (mmol/d)] (2)
Yplasma magnesium ¼ 2.543 þ [1.569  CCr (mL/s/1.73 m2)] (3)
If the values of Yfasting glucose and/or Yplasma magnesium calculated
using these equations are less than or equal to zero then the proba-
bility of the individual presenting hyperglycemia and/or hypomag-
nesemia, respectively, is higher. In contrast, if the derived parameters
are higher than zero there is greater probability of the individual
being normal. On the basis of the classification criteria embodied in
Eqs. (2) and (3), it is possible to infer that the patients were correctly
stratified according to fasting glucose and to plasma magnesium with
a predictive power of 78 and 59%, respectively (Table 3).
362 C.H. Sales et al. / Clinical Nutrition 30 (2011) 359e364

Table 1
Baseline characteristics of patients with type 2 diabetes distributed according to fasting glucose and plasma magnesium levels.

Characteristics Fasting glucose Plasma magnesium Overall mean

(n ¼ 51)
<6.11 mmol/L 6.11 mmol/L <0.75 mmol/L 0.75 mmol/L
(n ¼ 23) (n ¼ 28) (n ¼ 32) (n ¼ 19)
Age (y) 55.8  9.9 51.9  10.8 54.2  11.0 52.7  9.8 53.6  10.5
Duration of diabetes (y) 9.7  8.5 11.2  7.4 11.0  7.3 9.8  8.9 10.5  7.9
Duration of hypertension (y) 4.2  5.5 6.5  7.6 6.8  7.3 3.2  5.3 5.4  6.8
Systolic blood pressure (mmHg) 14.2  2.6 13.6  2.1 14.2  2.7 13.3  1.6 13.9  2.3
Diastolic blood pressure (mmHg) 8.9  2.0 8.6  1.1 8.8  1.8 8.6  1.1 8.7  1.6
Family history of diabetes (%) 65.2 85.7 81.2 68.4 76.5
Diagnosis of hypertension (%) 78.3 71.4 78.1 68.4 74.5
Medication users (%) 91.3 96.4 93.8 94.7 85.7
Performance of physical activity (%) 39.1 50.0 46.9 42.1 45.1
Smoking habits (%) 13.0 7.1 9.4 10.5 7.1
Weight (kg) 69.7  14.7 74.0  16.9 73.7  16.8 69.5  14.4 72.1  15.9
Height (m) 1.57  0.09 1.55  0.08 1.55  0.08 1.58  0.10 1.56  0.09
Body mass index (kg/m2) 28.3  5.08 30.4  5.41 30.5  5.6 27.7  4.36 29.4  5.3
Normal weight (%)a 30.4 7.1 9.4 31.6 17.6
Pre-obesity (%)b 34.8 46.4 46.9 31.6 41.2
Obesity class I (%)c 21.7 28.6 25.0 26.3 25.5
Obesity class II (%)d 13.0 10.7 12.5 10.5 11.8
Obesity class III (%)e 0.0 7.1 6.2 0.0 3.9
Waist circumference (cm) 96.4  11.0 100.6  12.1 100.5  11.8 95.8  11.3 98.7  11.7
Normal (%) 34.8 25.0 87.5 47.4 29.4
Elevated (%)f 65.2 75.0 12.5 52.6 70.6

The results are expressed in the form of mean  standard deviation or %, as appropriate.
BMI classifications3: a18.5e24.9 kg/m2, b25.0e29.9 kg/m2, c30.0e34.9 kg/m2, d35.0e39.9 kg/m2, e40.0 kg/m2.
Waist circumference classification3: felevated levels e males >102 cm, females >88 cm.

4. Discussion elimination of the mineral. However, individuals with hypomag-

The results presented here show that magnesium intake by the
study population was inadequate and that a high percentage of
individuals presented alterations in the status of this mineral. The
concentration of plasma magnesium was inversely correlated with
fasting and 2-h postprandial glucose, whereas the level of urine
magnesium was directly associated with fasting glucose. Moreover, as
shown by discriminant analyses, the glycemic levels of patients with
type 2 diabetes were influenced by magnesium while the concen-
trations of plasma magnesium were not influenced by glycemia.
CCr was the best marker of global kidney function and indicated
that the decline in kidney function in subjects with type 2 diabetes
may be associated with alterations in magnesium status. It might
be assumed that lower CCr values would imply an increased
concentration of circulating magnesium caused by the reduced
nesemia exhibited CCr values that were significantly smaller than
those of subjects presenting normal magnesemia, together with
microalbuminuria, proteinuria and increased serum creatinine
(Table 2).
According to Dewitte et al.27 a reduction in CCr in patients with
diabetes is not accompanied by an increase in serum ionized and
total magnesium as would be the case in normal individuals
without associated diseases. Moreover, it has been demonstrated
that a decline in kidney function in patients with type 2 diabetes is
accompanied by a decrease in serum magnesium,15,16,28 even when
such dysfunctions are not severe.17 Thus, small alterations in the
kidney function may induce changes in magnesium status.
Although osmotic diuresis can result in magnesium deficiency,
none of the four patients with polyuria (volume of urine >2500 mL/d)
presented increased excretion of magnesium. On the contrary, in one

Table 2
Dietary and biochemical characteristics of patients with type 2 diabetes distributed according to fasting glucose and plasma magnesium levels.

Parameters Fasting glucose p Plasma magnesium p Overall mean

(n ¼ 51)
<6.11 mmol/L 6.11 mmol/L <0.75 mmol/L 0.75 mmol/L
(n ¼ 23) (n ¼ 28) (n ¼ 32) (n ¼ 19)
Kidney function
Microalbuminuria (mg/d)a 45.7  76.7 (11.1) 48.4  76.9 (16.8) 0.248 50.9  80.4 (17.8) 41.0  69.6 (11.1) 0.235 47.2  76.0 (14.2)
Proteinuria (g/d)a 0.19  0.41 (0.09) 0.21  0.26 (0.14) 0.182 0.25  0.42 (0.13) 0.13  0.11 (0.09) 0.399 0.20  0.34 (0.11)
Creatinine clearance (mL/s/1.73 m2) 1.56  0.71 (1.50) 1.68  0.62 (1.62) 0.551 1.47  0.59 (1.39) 1.89  0.69 (1.67) 0.023 1.63  0.66 (1.51)
Urine volume (mL/d) 1612  618 (1750) 1733  632 (1701) 0.495 1504  528 (1505) 1973  671 (1835) 0.008 1678  623 (1712)

Blood glucose control

Fasting glucose (mmol/L) 5.0  0.7 (5.3) 10.6  3.3 (9.9) 0.000 8.7  3.5 (7.7) 7.0  3.8 (5.6) 0.104 8.1  3.7 (7.2)
2-h postprandial glucose (mmol/L) 7.3  2.8 (6.9) 14.2  4.4 (14.5) 0.000 12.0  4.9 (11.4) 9.5  5.1 (9.1) 0.085 11.1  5.1 (10.2)
Glycated hemoglobin (%) 9.9  2.4 (9.7) 12.6  2.9 (12.7) 0.001 12.0  2.7 (11.9) 10.3  3.2 (9.9) 0.047 11.4  3.0 (11.0)

Magnesium status and intake

Plasma magnesium (mmol/L) 0.75  0.07 (0.77) 0.68  0.07 (0.68) 0.001 0.66  0.06 (0.67) 0.79  0.02 (0.79) 0.000 0.71  0.08 (0.72)
Erythrocyte magnesium (mmol/L) 2.01  0.17 (2.05) 1.85  0.25 (1.83) 0.009 1.85  0.23 (1.85) 2.05  0.18 (2.05) 0.003 1.92  0.23 (2.00)
Urine magnesium (mmol/d) 2.48  1.34 (2.22) 3.06  1.61 (2.56) 0.180 2.56  1.63 (2.24) 3.20  1.22 (3.60) 0.147 2.80  1.51 (2.42)
Dietary magnesiumb (mmol/d) 9.80  2.03 (9.50) 9.02  1.44 (9.05) 0.116 9.49  1.89 (9.12) 9.19  1.53 (9.47) 0.562 9.37  1.76 (9.25)

The results are expressed in the form of mean  standard deviation with the median value shown in parenthesis.
a
Data not normally distributed (p < 0.05) according to the KolmogoroveSmirnov test.
b
Data adjusted according to the individual energy intake and intra-individual variability. The estimated average requirements23 are: a) for individuals 19e30 years old,
13.6 mmol/d (males), 10.5 mmol/d (females); b) for individuals >30 years old, 14.4 mmol/d (males), 10.9 mmol/d (females).
 C.H. Sales et al. / Clinical Nutrition 30 (2011) 359e364
Fig. 1. Probability of adequacy of magnesium intake by the studied patients with type 2
diabetes determined according to the methodology suggested by the Dietary Reference
Intakes of the United States Institute of Medicine.23,24 Adjusted values and reference
percentiles (P) are shown (EAR, estimated average requirement; RDA, recommended
dietary allowance; D, difference between the mean of magnesium intake and the
median of requirement; SDD, standard deviation of D).
of the subjects the excretion of magnesium in the urine and the
concentration of plasma magnesium were quite low.
Apparently, the kidney is the key organ in maintaining magnesium
homeostasis. It is possible, therefore, that the hypomagnesuria detec-
ted in the majority of patients resulted from the increased necessity of
reabsorbing magnesium in order to compensate for the low concen-
trations in the plasma and, possibly, low dietetic intake.23,29
It is important to emphasize that magnesium is an intracellular
ion, therefore concentration of magnesium in plasma or serum
must be considered to be a poor indicator of body magnesium
content1 even though this parameter has been used in many
studies. The magnesium plasma pool is substantially interchange-
able and normal concentrations are maintained by means of
reduction in urine excretion. However, plasma magnesium is the
second best indicator of the body status of this mineral after urine
magnesium,29 therefore its evaluation must not be disregarded.
Moreover, as demonstrated by the multivariate analyses presented
here, plasma magnesium together with urine and dietetic magne-
sium are the main predictors of hyperglycemia.
Within the patients with diabetes evaluated, 63% presented low
concentrations of plasma magnesium indicating alterations in the
compartmentalization of this mineral. The frequency established in
the present study was higher than the 13.5e47.7% range recorded
by Pham et al.17 but smaller than the value of 75% previously
reported by Lima et al.18
Table 3
Predicted group membership according to classification by fasting glucose and plasma magnesium.
Classifications Fasting glucosea
Normoglycemia
Normoglycemia n 19
% 82.6
Hyperglycemia n 7
% 25.0
Hypomagnesemia n e
% 65.6
Normomagnesemia n e
% 52.6
a
78.4% of the original grouped cases correctly classified according to the discriminant function generated (Eq. (2)).
b
58.8% of the original grouped cases correctly classified according to the discriminant function generated (Eq. (3)).
363
Although erythrocyte magnesium exerted no influence on the
glycemic control of the patients studied, some individuals presented
alterations in this compartment (Table 2) concomitant with changes
in the other magnesium parameters evaluated. Such indications
suggest that these subjects were at greater risk of developing co-
mplications and highlight the need for dietetic orientation regarding
adequate magnesium intake. It is important to emphasize, however,
that magnesium intake was inadequate in most (82%) of the subjects
studied, with the mean concentration value for the group
(6.28  1.16 mmol/1000 kcal) being lower than that observed in other
studies.12,14 The incidence of inadequate intake of magnesium in the
studied group was much higher than that observed in an investiga-
tion carried out in Australia13 in which only 26% of individuals pre-
sented magnesium intake levels below the EAR. Low magnesium
intake and reduced plasma magnesium concentrations may lead to
alterations in glycemic control because the distribution of the mineral
influences, and is influenced by, the secretion and action of insulin.
Within pancreatic b-cells, magnesium is essential for the activa-
tion of some phosphate-dependent enzymes that take part in the
glycolytic pathway and Krebs cycle, as well as for the transcription of
nuclear protein factors that are required for the release of insulin.
Moreover, the function of insulin is totally dependent on magnesium
since this mineral is responsible for the activation of the b-subunit of
the tyrosine kinase domain of the insulin receptor, and for the stim-
ulation of proteins and substrates in the insulin-signaling cascade.2,11
Since magnesium is essential owing to its involvement in the
magnesiumeATP complex that takes part in all phosphate transfer
reactions that use and supply energy,4 it is not surprising that defi-
ciency of this mineral is implicated in the impairment of metabolic
control, as confirmed in the present study, and for additional chronic
complications as previously described.15e19,27,28 However, some
limitations of the present study must to be taken into consideration.
Firstly, evaluation of dietary intake is susceptible to random and
systematic errors, although in the present study the 24-h recalls were
applied by a trained nutritionist with the aid of an album containing
images of food and utensils in order to improve the accuracy of dietary
description by patients. Additionally, the data obtained were subse-
quently adjusted on the basis of energy and intra-individual variation.
Secondly, the method employed for the evaluation of HbA1 only
quantified total HbA1. Finally, a cross-sectional study not only limits
the ability to make causal inferences, but is also susceptible to bias of
prevalence/incidence. In the present study, stepwise discriminant
analysis was employed with the purpose of overcoming such
constraints and to raise a greater number of inferences, including the
generation of equations that could be applied in clinical practice.
In conclusion, the results presented herein show that impaired
kidney function may lead to hypomagnesemia and that this condi-
tion, together with low magnesium intake and excretion, can induce
the augmentation of glucose in the blood. Long-term hyperglycemia
Plasma magnesiumb
Hyperglycemia Hypomagnesemia Normomagnesemia
4 e e
17.4
21 e e
75.0
e 21 11
34.4
e 10 9
47.4
364 C.H. Sales et al. / Clinical Nutrition 30 (2011) 359e364
in patients with type 2 diabetes increases the risk of chronic
complications, such as nephropathy, which may exacerbate hypo-
magnesemia and aggravate the clinical conditions. Hence, the initial
hypothesis that an adequate magnesium intake is essential for
subjects with type 2 diabetes was confirmed. Analysis of magne-
sium status in the routine assessment of such patients, or at least for
those who are not able to reach the desired glycemic standards,
would be useful in evaluating the risks relating to chronic compli-
cations in diabetes. Such strategy may help the management of type 2
diabetes and reduce the risk of long-term severe complications.
Sources of funding
The study was financially supported by the Fundação de Amparo
à Pesquisa do Estado de São Paulo (FAPESP; process #2006/05601-
9), the Conselho Nacional de Desenvolvimento Científico e Tecno-
lógico (CNPq; process #135858/2006-2, 480487/2009-0), the
Ministério da Saúde do Brasil and the Fundação de Apoio à Pesquisa
do Estado do Rio Grande do Norte (FAPERN; process PPSUS/2007
#175_19969779).
Statement of authorship
We hereby certify that it is an original publication. This manu-
script is not currently under consideration or in press elsewhere.
The contributions of the authors to the manuscript are as follows.
C.H.S. contributed to the conception and design of the study, acqui-
sition of data, samples analyses, statistical analysis, analysis and data
interpretation, drafting the article and revising it critically for
important intellectual content. L.F.C.P. contributed to the conception
and design of the study and its coordination, analysis and interpre-
tation of data, drafting the article and revising it critically for impor-
tant intellectual content. J.G.L. contributed to the design of the study,
acquisition of data, analysis and interpretation of data, revising the
article critically for important intellectual content. T.M.A.M.L.
contributed to the design of the study, samples analyses, analysis and
interpretation of data, revising the article critically for important
intellectual content. C.C. contributed to the conception and design of
the study and its coordination, analysis and data interpretation,
drafting the article and revising it critically for important intellectual
content. All authors read and approved this submitted version.
Conflict of interest
None of the authors do have any potential conflicts of interest
that could inappropriately influence in this work.
Acknowledgments
The authors are grateful to the director of the HUOL and all of
the participants for their valuable collaboration. Particular thanks
go to Prof. Maria das Graças Almeida, Ana Rachel F. Barbosa and
Luzia L. S. Fernandes, from the Department of Clinical and Toxico-
logical Analyses, Federal University of Rio Grande do Norte, for
technical assistance with laboratory analyses, and to Prof. Julia
Maria P. Soler, from the Institute of Mathematics and Statistics,
University of São Paulo, for help with statistical interpretation.
References
1. Topf JM, Murray PT. Hypomagnesemia and hypermagnesemia. Rev Endocr
Metab Disord 2003;4:195e206.
2. Takaya J, Higashino H, Kobayashi Y. Intracellular magnesium and insulin
resistance. Magnes Res 2004;17:126e36.
3. American Diabetes Association. Nutrition recommendations and interventions
for diabetes. Diabetes Care 2008;31:S61e78.
4. Sales CH, Pedrosa LFC. Magnesium and diabetes mellitus: their relation. Clin
Nutr 2006;25:554e62.
5. Kao WHL, Folsom AR, Javier Nieto F, Mo J-P, Watson RL, Brancati FL. Serum and
dietary magnesium and the risk for type 2 diabetes mellitus: the atheroscle-
rosis risk in communities study. Arch Intern Med 1999;159:2151e9.
6. Lopez-Riadura R, Willett WC, Rimm EB, Liu S, Stampfer MJ, Manson JE, et al.
Magnesium intake and risk of type 2 diabetes in men and women. Diabetes
Care 2004;27:134e40.
7. van Dam RM, Hu FB, Rosenberg L, Krishnan S, Palmer JR. Dietary calcium and
magnesium, major food sources, and risk of type 2 diabetes in U.S. black
women. Diabetes Care 2006;29:2238e43.
8. Villegas R, Gao YT, Dai GQ, Yang G, Cai H, Li H, et al. Dietary calcium and
magnesium intakes and the risk of type 2 diabetes: the Shanghai women’s
health study. Am J Clin Nutr 2009;89:1059e67.
9. Kim DJ, Xun P, Liu K, Loria C, Yokota K, Jacobs Jr DR, et al. Magnesium intake in
relation to systemic inflammation, insulin resistance, and the incidence of
diabetes. Diabetes Care; 2010; doi:10.2337/dc10-0994.
10. Nanri A, Mizoue T, Noda M, Takahashi Y, Kirii K, Inoue M, et al. Magnesium
intake and type II diabetes in Japanese men and women: the Japan Public
Health Center-based prospective study. Eur J Clin Nutr 2010;64:1244e7.
11. Suárez A, Pulido N, Casla A, Casanova B, Arrieta FJ, Rovira A. Impaired tyrosine-
kinase activity of muscle insulin receptors from hypomagnesaemic rats. Dia-
betologia 1995;38:1262e70.
12. Mayer-Davis EJ, Nichols M, Liese AD, Bell RA, Dabelea DM, Johansen JM, ,
et alSEARCH for Diabetes in Youth Study Group. Dietary intake among youth
with diabetes: the SEARCH for diabetes in youth study. J Am Diet Assoc
2006;106:689e97.
13. Pitt S, May K, Colman P, Wraight P. Deficiencies in nutritional intake in patients
admitted with diabetes-related foot complications. Nutr Diet 2007;64:186e91.
14. Ma J, Folsom AR, Melnick SL, Eckfeldt JH, Sharrett AR, Nabulsi AA, et al. Asso-
ciations of serum and dietary magnesium with cardiovascular disease, hyper-
tension, diabetes, insulin, and carotid arterial wall thickness: the ARIC study.
J Clin Epidemiol 1995;48:927e40.
15. Pham PC, Pham PM, Pham PA, Pham SV, Pham HV, Miller JM, et al. Lower serum
magnesium levels are associated with more rapid decline of renal function in
patients with diabetes mellitus type 2. Clin Nephrol 2005;63:429e36.
16. Prabodh S, Prakash DSRS, Sudhakar G, Chowdary NVS, 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; 2010; doi:10.1007/s12011-
010-8750-x.
17. Pham PC, Pham PM, Pham PT, Pham SV, Pham PA, Pham PT. The link between
lower serum magnesium and kidney function in patients with diabetes mel-
litus type 2 deserves a closer look. Clin Nephrol 2009;71:375e9.
18. Lima ML, Pousada J, Barbosa C, Cruz T. Deficiência de magnésio e resistência à
insulina em pacientes com diabetes mellitus tipo 2. Arq Bras Endocrinol Metabol
2005;49:959e63.
19. Paula FJA, Lanna CMM, Shuhama T, Foss MC. Effect of metabolic control on
parathyroid hormone secretion in diabetic patients. Braz J Med Biol Res
2001;34:1139e45.
20. Sociedade Brasileira de Diabetes. Tratamento e Acompanhamento do Diabetes
Mellitus. Diretrizes da Sociedade Brasileira de Diabetes. Rio de Janeiro: Dia-
graphic; 2007.
21. Sales CH, Setaro L, Rocha VS, Colli C. Magnésio plasmático e urinário: validação
de métodos de análise. Braz J Pharm Sci. 2008;44:57.
22. Tietz NW, editor. Clinical guide to laboratory tests. 3rd ed. Philadelphia: WB
Saunders; 1995.
23. Food and Nutrition Board, Institute of Medicine. Magnesium. In: Food and
Nutrition Board, Institute of Medicine, editor. Dietary reference intakes for
calcium, phosphorus, magnesium, vitamin D, and fluoride. Washington: National
Academy Press; 1997. p. 190e249.
24. Food and Nutrition Board, Institute of Medicine. Using dietary reference
intakes for nutrient assessment of individuals. In: Food and Nutrition Board,
Institute of Medicine, editor. Dietary reference intakes: applications in dietary
assessment. Washington: National Academy Press; 2000. p. 45e69.
25. Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in epide-
miologic studies. Am J Clin Nutr 1997;65:1220Se8S.
26. National Research Council, Coordinating Committee on Evaluation of Food
Consumption Surveys, Subcommittee on Criteria for Dietary Evaluation.
Nutrient adequacy: assessment using food consumption surveys. Washington:
National Academy Press; 1986.
27. Dewitte K, Dhondt A, Giri M, Stöckl D, Rottiers R, Lameire N, et al. Differences in
serum ionized and total magnesium values during chronic renal failure
between nondiabetic and diabetic patients. A cross-sectional study. Diabetes
Care 2004;27:2503e5.
28. Corsonello A, Ientile R, Buemi M, Cucinotta D, Mauro VN, Macaione S, et al.
Serum ionized magnesium levels in type 2 diabetic patients with micro-
albuminuria or clinical proteinuria. Am J Nephrol 2000;20:187e92.
29. Sabatier M, Pont F, Arnaud MJ, Turnlund JR. A compartmental model of
magnesium metabolism in healthy men based on two stable isotope tracers.
Am J Physiol Regul Integr Comp Physiol 2003;285:R656e663.