You are on page 1of 8

Effect of oral magnesium supplementation on physical performance

in healthy elderly women involved in a weekly exercise program:

a randomized controlled trial1–3
Nicola Veronese, Linda Berton, Sara Carraro, Francesco Bolzetta, Marina De Rui, Egle Perissinotto, Elena Debora Toffanello,
Giulia Bano, Simona Pizzato, Fabrizia Miotto, Alessandra Coin, Enzo Manzato, and Giuseppe Sergi

ABSTRACT is also associated with an increased inflammatory state, muscle

Background: Magnesium deficiency is associated with poor phys- cell alterations attributable to increased oxidative stress, and
ical performance, but no trials are available on how magnesium impaired intracellular calcium homeostasis (7). All these factors
supplementation affects elderly people’s physical performance. negatively affect muscle mass and function and could exacerbate
Objective: The aim of our study was to investigate whether 12 wk the sarcopenia typical of old age.
of oral magnesium supplementation can improve physical perfor- Elderly people are particularly susceptible to magnesium
mance in healthy elderly women. deficiency for several reasons, including an inadequate dietary
Design: In a parallel-group, randomized controlled trial, 139 intake, a less efficient magnesium absorption, and greater losses
healthy women (mean 6 SD age: 71.5 6 5.2 y) attending a mild in stools and urine (2). Elderly women may be more susceptible
fitness program were randomly allocated to a treatment group (300 mg to magnesium deficiencies than men, partly because they are
Mg/d; n = 62) or a control group (no placebo or intervention; n = 77) more likely to have osteoporosis, which limits the exchange of
by using a computer-generated randomization sequence, and re- magnesium between bone and blood (8). Older women also have
searchers were blinded to their grouping. After assessment at base- a low magnesium intake in their diet: in Europe, 77% of women
line and again after 12 wk, the primary outcome was a change in the older than 65 y have a dietary intake below the Recommended
Short Physical Performance Battery (SPPB); secondary outcomes Dietary Allowance (RDA),4 and 23% ingest less than two-thirds
were changes in peak torque isometric and isokinetic strength of the of the RDA (2, 9).
lower limbs and handgrip strength. It has been widely accepted that magnesium has a positive
Results: A total of 124 participants allocated to the treatment (n = 53) effect on muscle function, but studies on the efficacy of mag-
or control (n = 71) group were considered in the final analysis. At
nesium supplementation in young people have generated con-
baseline, the SPPB scores did not differ between the 2 groups. After
trasting results (10–15). Differences in magnesium status may be
12 wk, the treated group had a significantly better total SPPB score
the reason for these different findings, because some researchers
(D = 0.41 6 0.24 points; P = 0.03), chair stand times (D = 21.31 6
have suggested that magnesium supplementation might only
0.33 s; P , 0.0001), and 4-m walking speeds (D = 0.14 6 0.03 m/s;
help magnesium-deficient people (16), and many studies did not
P = 0.006) than did the control group. These findings were more
assess magnesium status (or they did so using scarcely sensitive
evident in participants with a magnesium dietary intake lower than
methods) before providing supplementation. Finally, to the best
the Recommended Dietary Allowance. No significant differences
emerged for the secondary outcomes investigated, and no serious of our knowledge, no such investigations appear to have been
adverse effects were reported. conducted in older people.
Conclusions: Daily magnesium oxide supplementation for 12 wk
seems to improve physical performance in healthy elderly women. 1
From the Department of Medicine–DIMED, Geriatrics Section, Univer-
These findings suggest a role for magnesium supplementation in pre- sity of Padova, Padova, Italy (NV, LB, SC, FB, MDR, EDT, GB, SP, FM,
venting or delaying the age-related decline in physical performance. AC, EM, and GS), and the Department of Cardiac, Thoracic and Vascular
This trial was registered at as NCT01971424. Sciences, Biostatistics, Epidemiology and Public Health Unit, University of
Am J Clin Nutr 2014;100:974–81. Padova, Padova, Italy (EP).
No funding was used in the design, implementation, analysis, or inter-
pretation of the data. Magnesium samples were provided free of charge by
INTRODUCTION Sanofi-Aventis.
Magnesium has a fundamental role in muscle function and is Address correspondence to N Veronese, Department of Medicine–
essential to energy metabolism, transmembrane transport, and DIMED, Geriatrics Division, University of Padova, Via Giustiniani, 2
35128 Padova, Italy. E-mail:
muscle relaxation and contraction (1–4). Magnesium deficiency 4
Abbreviations used: PASE, Physical Activity Scale for the Elderly; PT,
has proved capable of impairing exercise capacity and reducing peak torque; RDA, Recommended Dietary Allowance; SPPB, Short Physical
physical performance (5). A large cross-sectional study on older Performance Battery; 25(OH)D, 25-hydroxyvitamin D.
people showed a strong association between serum magnesium Received November 18, 2013. Accepted for publication June 9, 2014.
and several muscle performance tests (6). Magnesium depletion First published online July 9, 2014; doi: 10.3945/ajcn.113.080168.

974 Am J Clin Nutr 2014;100:974–81. Printed in USA. Ó 2014 American Society for Nutrition

Downloaded from

by guest
on 25 April 2018
Our hypothesis is that oral magnesium supplementation could supplemented with 800 IU cholecalciferol/d orally, and their
improve physical performance and muscle function in old concentrations were remeasured after 1 mo. Those whose 25(OH)D
women, and the aim of our study was thus to test whether 12 wk concentrations remained ,50 nmol/L after supplementation
of daily oral magnesium supplementation (300 mg Mg) could (n = 2) were ruled out before the randomization.
improve muscle strength and physical performance in a group of During the follow-up, all participants continued their fitness
healthy elderly women. program and their attendance was monitored and recorded bi-
weekly in their records. The use of new supplements or medi-
SUBJECTS AND METHODS cations during the study period was not permitted and was
considered a reason for exclusion, but none of the women in the
Participants sample met this exclusion criterion.
This study was conducted at the Geriatrics Department of The study was designed in accordance with the Helsinki
Padova University between January 2012 and March 2013. Declaration and was approved by the local Ethical Committee (no.
Women older than 65 y were recruited on a voluntary basis from 491/2011). All participants were fully informed about the nature,
among elderly people attending a twice-weekly mild fitness purpose, procedures, and risks of the study and gave their informed
program at public gyms in Padova, Italy. consent. Our trial complied with the Consolidated Standards of
Candidates were excluded if they had evidence of renal failure, Reporting Trials statement for randomized trials (Figure 1) (18).
chronic or acute infection, a history or evidence of malignancy in
the past 5 y (except for nonmelanoma skin neoplasia), significant
cardiovascular or pulmonary diseases, uncontrolled metabolic Randomization, intervention, and allocation
disease (diabetes, anemia, or thyroid disease), or electrolyte Volunteers were randomly assigned to 1 of 2 experimental
abnormalities or if they used any drugs or supplements that might groups by using a computer-generated sequence of 139 nonunique,
interfere with magnesium metabolism (magnesium, potassium, unsorted numbers with a range from 1 to 2 representing the groups.
calcium, digitalis, or proton pump inhibitors). During a screening The main investigator (NV) kept the allocation sequence confi-
visit, their healthy condition was established by trained medical dential and assigned the women to one or the other group.
personnel on the strength of their medical history, a clinical Participants received 12 wk of supplementation with 900 mg/d of
examination, and routine biochemical tests (eg, measurement of oral magnesium oxide corresponding to 300 mg bioavailable
glucose and electrolyte concentrations, renal and liver function magnesium [Easymag (Sanofi-Aventis); treatment group] or no
tests, and a complete blood count). Given the close relation treatment (control group). This dose of magnesium was chosen
between 25-hydroxyvitamin D [25(OH)D] and physical perfor- because it corresponds to 94% of the RDA (19), and a once-daily
mance in elderly people (17), individuals with 25(OH)D con- administration was adopted to improve the participants’ compliance
centrations ,25 nmol/L (n = 14) at the screening visit were with the supplementation. The supplement was distributed to

FIGURE 1. Consolidated Standards of Reporting Trials diagram showing sample sizes at each stage of the study. 25(OH)D, 25-hydroxyvitamin D.

Downloaded from

by guest
on 25 April 2018

participants in sachets (daily doses of supplement) contained in We considered the following tertiary outcomes:
cartons (monthly supplies). Biweekly phone calls were made by
1) Blood and urine tests: venous blood samples were collected
the research coordinator (LB) to record compliance with the
after the subjects fasted overnight. Biohumoral tests were
treatment and any adverse events and to monitor whether par-
completed at our Department of Laboratory Medicine,
ticipants were attending their fitness program. Their attendance
which belongs to the International Federation of Clinical
was also recorded weekly in registers kept at the gyms.
Chemistry and Laboratory Medicine. Magnesium was
At baseline, a trained dietitian asked the control group to avoid
measured in serum and in 24-h urine samples by using
modifying their usual diet, and they received the same follow-up
a standard colorimetric enzymatic method that consists of
phone calls as the treated group. The 2 groups were followed up by 2
a coupled enzyme assay, which resulted in a colorimetric
independent teams of physicians who conducted physical and body-
(450 nm) product proportional to the magnesium content.
composition tests at baseline and again after 12 wk. The same
The intra- and interassay CVs for serum magnesium were
professional on each team performed the same test at both time
1.8% and 1.4%, respectively, at a concentration of 0.81
points. To ensure the success of the masking procedure, these 2 teams
mmol/L, whereas the corresponding values for 24-h urinary
of physicians did not know whether a given participant belonged to
magnesium were 0.6% and 1.6% at a concentration of 2.45
the treatment or the control group (the participants had been trained
mmol/L. On the basis of quality-control materials (Liquichek
not to reveal this information to the personnel performing the tests).
Unassayed Chemistry Control; Bio-Rad), the interassay CVs
for magnesium covered a range of concentrations from 1.8%
Outcomes to 5.1%. Serum 25(OH)D concentrations were measured by
All tests were performed at baseline and again after 12 wk. Our radioimmunoassay (DiaSorin); the intraassay and interas-
primary outcome was a change in the Short Physical Performance say CVs for 25(OH)D were 7.9% and 8.2%, respectively.
Battery (SPPB) (20): Serum intact parathormone concentrations were measured
by using a 2-step immunoradiometric assay (PTH-S; N-tact
1) The SPPB consists of 3 objective physical function tests: PTH SP; DiaSorin): the intraassay and interassay CVs for
4-m gait speed, repeated chair stands, and standing balance parathormone were 3.0% and 5.5%, respectively. For both
in increasingly challenging positions. Walking speed was procedures, the standards, controls, and samples were all
calculated as the best performance achieved in 2 walks at analyzed in duplicate, and the samples were reanalyzed if
the participant’s usual pace along a corridor 4 m long. For the CVs of the duplicates were .10%.
the chair stands test, the participants were asked to rise 5
times from a seated position as quickly as possible with 2) Body composition: body weight and height were measured by
their hands folded across their chest. For the standing bal- trained staff. Body composition was assessed by dual-energy
ance tests, participants were asked to stand in 3 increas- X-ray absorptiometry with a fan-beam technology (Hologic
ingly difficult positions (with their feet side by side and in Discovery A) (21). The Appendicular Skeletal Muscle Mass
semitandem and full-tandem positions) for 10 s each. Each Index, ie, the ratio of appendicular skeletal muscle mass to
test was scored from 0 (worst) to 4 (best), and the scores height (in kg/m2), was considered to be the best indicator of
from all 3 tests were combined to obtain a composite score functional muscle mass in elderly people (22, 23).
of 0 to 12—higher scores reflect better physical function. 3) Physical and dietary assessment: physical activity was mea-
Our secondary outcomes were variables of muscle strength: sured by using the Physical Activity Scale for the Elderly
(PASE), which is designed to assess physical activities in
1) Isometric knee extension torque and isokinetic (flexion and elderly adults within a 1-wk time frame (24). The scale covers
extension) strength were tested on the dominant side by 12 different activities, such as walking, sports, and house-
using the dynamometer chair (Easytech s.r.l.). The partici- work, and is scored from 0 to 400 and more (no maximum
pants were positioned upright with straps to fix their hips to score has been defined). The type of activity and its fre-
the chair. For each of the 3 measurements, participants were quency, and the time spent on it, were recorded and the PASE
asked to reach their maximal voluntary contraction. Three to scores were calculated as explained in the PASE manual (25).
5 s after reaching their maximum effort, they were asked to At baseline, a trained dietitian conducted a dietary assessment
stop the contraction. Each measurement was repeated 3 using a modified dietary assessment method involving an
times, and patients rested for 2 to 3 min between trials. estimated 3-d record and a questionnaire on the frequency
The highest-peak torque (PT) was used for the analysis. with which participants usually ate certain foods; data from
2) Handgrip strength on the dominant side was measured by the previous month were used as reference (26, 27). The usual
using DynEx electronic hand dynamometers. The partici- food intake was converted into macronutrients and micronu-
pants were seated in a standard armchair with their shoul- trients by using a national food-composition table (28).
der adducted and neutrally rotated, their elbow flexed at
908, and their forearm and wrist in a natural position. They Sample size and statistical analyses
were asked to grip the dynamometer smoothly, progressing The required sample size was calculated from the difference of
up to their maximal strength in response to a voice com- 0.28 points in total SPPB scores between the treated and control
mand, without any wrenching or jerking motion. Three groups after 12 wk (29). With an SD of 0.5 points, the number of
measurements were obtained on the dominant side with participants estimated to be needed in each group was 50 for an
a 1-min rest between trials, and the highest measurement 80% power and a = 0.05. This sample size was also able to
was used in our analyses. identify a difference of 4.0 6 5.6 kg in handgrip strength as

Downloaded from

by guest
on 25 April 2018
significant, which had been found to be clinically significant in was 27.40 6 4.18. During the follow-up, 10 participants were lost:
other studies (this applied to n = 31 participants per group) (30). these women were similar to the participants included in the anal-
Finally, because no unequivocal data are available for changes in ysis in terms of age, baseline physical performance and strength,
isometric and isokinetic strength in old age, for the purposes of blood and serum tests, body composition, and levels of physical
this study we assumed that a difference of 5 6 5% for each activity (details not shown).
variable was clinically significant (which applied to n = 16 par- After the treated participants were separated from the control
ticipants per group). subjects, no significant differences emerged for age (71.8 6 5.0
Baseline characteristics were compared between the treatment compared with 71.3 6 5.4 y, respectively; P = 0.62), daily calorie
and control groups by using independent t tests, chi-square tests, intake (1454.76 6 423.34 compared with 1526.57 6 333.56 kcal;
or Fisher’s exact test, as appropriate. Paired t tests were used for P = 0.29), or baseline magnesium intake (394.00 6 230.13 com-
within-group comparisons of baseline and 12-wk data, and pared with 372.05 6 163.07 mg/d; P = 0.56). There were 24 par-
changes were calculated as the difference between the 2 values ticipants with a magnesium intake below the RDA in the treated
(D). Comparisons between data for the groups at 12 wk were group and 30 in the control group (P = 0.86). The number of
computed by using a generalized linear model, adjusted for the participants in the 2 groups taking vitamin D supplementation was
baseline value of the corresponding test. In the ancillary analysis, also comparable: 23 in the treated group and 35 in the control group
we compared the changes in SPPB scores at 12 wk using a strati- (P = 0.59).
fication for baseline SPPB total score and dietary magnesium in-
take. P for interaction was calculated by using a generalized linear Compliance
model with the baseline SPPB score categorized as #10 or 10 and
On the basis of sachet and carton counts and specific questions
the dietary magnesium intake as below or reaching the corre-
posed during the study and at the follow-up visit, the percentage
spondent RDA. Pearson’s correlations were run on the combined
of prescribed magnesium doses ingested by the treatment group
data from both groups and were used to identify associations be-
was 90 6 0.5%.
tween serum indicators and primary outcome items. Significance
was accepted if P # 0.05, and all tests were 2-tailed. All analyses
Primary outcome
were performed by using SPSS 21.0 for Windows (SPPS Inc).
As shown in Table 1, no differences in total SPPB, single item
RESULTS scores, gait speed, or chair stand times were found between the 2
groups at baseline. When the variations in SPPB outcomes be-
Participants tween the 2 groups were compared, the treated group showed
The mean (6SD) age of the sample as a whole was 71.5 6 5.2 y, a significantly greater improvement than did the control group
the mean weight was 66.3 6 10.7 kg, and the mean BMI (in kg/m2) for the SPPB walking item (D = 0.34 6 0.10 points; P = 0.01),
Mean baseline characteristics and outcomes at follow-up, by group1
Magnesium oxide group (n = 53) Control group (n = 71)
Measure Baseline3 12 wk Change4 Baseline3 12 wk Change4 change (12 wk)2

Primary outcomes
SPPB-balance 3.62 6 0.66 3.58 6 0.77 20.04 6 0.78 3.55 6 0.75 3.76 6 0.57 0.21 6 0.77 20.20 6 0.11
SPPB-walking 3.13 6 0.68 3.66 6 0.52 0.53 6 0.72*** 3.39 6 0.60 3.41 6 0.60 0.01 6 0.62 0.34 6 0.10*
SPPB-chair stands 3.42 6 0.80 3.58 6 0.77 0.17 6 0.73 3.15 6 0.84 3.23 6 0.87 0.07 6 0.70 0.20 6 0.12
SPPB-total score 10.13 6 1.44 10.83 6 1.64 0.70 6 1.51** 10.08 6 1.38 10.39 6 1.53 0.31 6 1.31* 0.41 6 0.24*
Chair stands time (s) 10.99 6 2.29 9.77 6 2.84 21.21 6 1.86*** 11.91 6 2.55 11.86 6 2.63 20.05 6 1.80 21.31 6 0.33***
Walking speed (m/s) 1.16 6 0.20 1.37 6 0.21 0.21 6 0.27*** 1.25 6 0.18 1.27 6 0.19 0.02 6 0.16 0.14 6 0.03***
Secondary outcomes
PT isokinetic flex (nm) 58.64 6 15.39 59.14 6 13.89 0.50 6 8.33 60.38 6 14.65 58.65 6 15.04 21.73 6 8.41 2.57 6 1.54
PT isokinetic extension (nm) 25.06 6 9.62 25.64 6 9.00 0.58 6 7.30 25.54 6 6.85 25.27 6 7.35 20.28 6 5.27 0.76 6 1.23
PT isometric (nm) 73.81 6 19.28 79.03 6 23.28 5.22 6 17.70 86.11 6 19.04 77.08 6 18.02 29.03 6 15.28 13.33 6 2.57
Handgrip (kg) 22.08 6 5.00 23.59 6 5.33 1.52 6 3.50 21.16 6 5.81 21.36 6 4.91 0.19 6 3.23 0.51 6 0.56
Tertiary outcomes
Serum magnesium (nmol/L) 0.83 6 0.04 0.86 6 0.05 0.03 6 0.04*** 0.83 6 0.06 0.84 6 0.06 0.009 6 0.04 0.02 6 0.007*
Urinary magnesium (mmol/24 h) 3.80 6 1.47 5.16 6 2.51 1.36 6 2.30*** 3.79 6 1.34 3.91 6 1.28 20.12 6 1.17 1.45 6 0.31***
25(OH)D (mmol/L) 51.60 6 27.92 62.30 6 35.89 10.70 6 23.49 67.38 6 31.68 73.87 6 31.52 6.49 6 19.62 2.07 6 3.95
Parathormone (ng/L) 27.85 6 9.86 28.26 6 10.19 20.42 6 5.07 28.55 6 9.23 30.09 6 10.57 1.55 6 6.30* 21.19 6 1.05
ASMMI (kg/m2) 6.22 6 0.76 6.26 6 0.72 0.03 6 0.16 6.76 6 0.96 6.74 6 0.91 20.02 6 0.22 0.06 6 0.08
PASE score 199.53 6 52.48 200.06 6 52.80 0.53 6 45.60 208.65 6 42.64 203.85 6 38.21 24.80 6 23.76 2.45 6 5.76
All values are means 6 SDs. *P , 0.05; **P , 0.001; ***P , 0.0001. ASMMI, Appendicular Skeletal Muscular Mass Index; PASE, Physical Activity
Scale for the Elderly; PT, peak torque; SPPB, Short Physical Performance Battery; 25(OH)D, 25-hydroxyvitamin D.
A generalized linear model was used for a between-group comparison of the data obtained after 12 wk, adjusted for the baseline value of the corresponding test.
There were no statistically significant differences between the groups in any baseline test.
Changes were calculated as the differences between data at baseline and at follow-up after 12 wk (within-group comparison) by using paired t tests.

Downloaded from

by guest
on 25 April 2018

the total SPPB score (D = 0.41 6 0.24 points; P = 0.03), the

change (12 wk)2 P-interaction3

chair stands time (D = 21.31 6 0.33 s; P , 0.0001), and




walking speed (D = 0.14 6 0.03 m/s; P = 0.0006) (Table 1).

20.10 6 1.20 21.52 6 0.30***


20.15 6 0.46 0.29 6 0.11*

0.09 6 0.05*
Ancillary analyses

20.04 6 0.33 20.10 6 0.15

0.00 6 0.48 0.14 6 0.09

0.39 6 0.20
Among other factors that might have contributed to our
findings, baseline total SPPB scores and dietary magnesium in-
take seem to be relevant. When our participants were stratified

20.19 6 0.68

0.01 6 0.14
by median baseline total SPPB scores, we found that the items in

the SPPB that improved significantly in the sample as a whole
remained the same, irrespective of this distinction (Table 2).

Control group (n = 27)

A generalized linear model was used for a between-group comparison of the data obtained after 12 wk, adjusted for the baseline value of the corresponding test.
Dividing the sample between those who reached the RDA of

0.18 6 0.27*** 1.38 6 0.12 1.39 6 0.0.15

3.89 6 0.32 3.85 6 0.46
3.85 6 0.36 3.70 6 0.46
3.78 6 0.42 3.78 6 0.42

11.52 6 0.51 11.33 6 0.68

13.00 6 2.38 12.99 6 2.52 20.01 6 2.10 21.22 6 0.52* 9.40 6 1.15 8.02 6 1.19 21.38 6 1.09*** 10.11 6 1.65 10.01 6 1.55
dietary magnesium and those who did not, we found similar

12 wk
differences between the treated and control groups in terms of the

Changes were calculated as the differences between data at baseline and at follow-up after 12 wk (within-group comparison) by using paired t tests.
changes in their chair stands time and walking speed (Table 3).
On the other hand, the SPPB total score was significantly higher

in the treated than in the control group, but only for participants

Baseline SPPB .10

with a magnesium intake below the RDA, and their better SPPB
score was probably the result of a significant improvement in the

0.29 6 0.15* 3.52 6 0.51 3.88 6 0.33 0.36 6 0.49**

SPPB walking item.

0.36 6 0.92* 20.27 6 0.16 3.84 6 0.37 3.72 6 0.68 20.12 6 0.73

0.26 6 0.20 4.00 6 0.00 4.00 6 0.00 0.00 6 0.00

0.32 6 0.75
Magnesium oxide group (n = 25)

Values at baseline were obtained by means of a general linear models procedure including the interaction term in each model.
Secondary outcome: PT isometric and isokinetic leg and
handgrip strength measurements

0.38 6 0.39 11.28 6 0.46 11.60 6 0.76

0.12 6 0.04* 1.27 6 0.17 1.45 6 0.20

All values are means 6 SDs. *P , 0.05; **P , 0.001; ***P , 0.0001. SPPB, Short Physical Performance Battery.
12 wk
At baseline, no differences in PT leg or handgrip strength were
found between the groups. After 12 wk, no significant within- or
between-group changes were found in PT isometric and iso-

kinetic strength. The changes in handgrip strength were also
similar between the groups (Table 1).

(12 wk)2

Tertiary outcomes

There were no statistically significant differences between the groups in any baseline test.
Blood and urine tests
No significant differences in serum and 24-h urinary magnesium
0.11 6 0.69
0.12 6 0.81

0.62 6 1.50

0.01 6 0.18
were found at baseline. A comparison of the changes for these 2

outcomes between the groups showed that both serum (D = 0.02 6

Control group (n = 44)

0.007 mmol/L; P = 0.03) and urinary magnesium (D = 1.45 6

Primary outcome in the treated and control groups by baseline SPPB score1

0.31 mmol/24 h; P , 0.0001) were significantly higher in the

3.34 6 0.86 3.70 6 0.63
0.68 6 0.86*** 3.11 6 0.54 3.23 6 0.61
2.77 6 0.80 2.89 6 0.90

9.20 6 0.93 9.82 6 1.62

0.23 6 0.27*** 1.18 6 0.17 1.19 6 0.17

12 wk

treated than in the control group. The changes from baseline in

25(OH)D and parathormone did not differ significantly between
the 2 groups (Table 1).
Baseline SPPB #10

Body composition
When the Appendicular Skeletal Muscle Mass Index was used
12.40 6 2.13 11.33 6 2.98 21.07 6 2.35*

as an indicator of functional lean mass, no difference was found

0.03 6 0.84

0.32 6 0.98

1.04 6 1.92
Magnesium oxide group (n = 28)

between the groups at baseline, and no significant differences

were found between the groups at 12 wk (Table 1). Similarly, no
significant differences in weight, height, BMI or total fat mass,
3.43 6 0.79 3.46 6 0.84
2.79 6 0.63 3.46 6 0.58
2.89 6 0.79 3.21 6 0.92

9.11 6 1.23 10.14 6 1.90

1.06 6 0.16 1.29 6 0.17

and fat-free mass emerged between the groups at baseline, and

12 wk

the situation remained the same after 12 wk (details not shown).

Physical assessment

No significant between-groups differences in the levels of

physical activity, based on the PASE score, were found at baseline
(P = 0.68) or at the follow-up (P = 0.67) visits (Table 1). Finally,
speed (m/s)

Chair stands

the number of participants who attended $80% of their fitness

times (s)




program meetings was high and was similar between the 2

groups (45 in the treated and 60 in the control group; P = 0.78).

Downloaded from

by guest
on 25 April 2018

change (12 wk)2 P-interaction3





Pearson’s correlations were used to determine whether the
changes in serum or urinary magnesium were associated with
changes in primary outcomes. Increases in serum magnesium

0.13 6 0.04**

20.96 6 0.42*
20.35 6 0.16
0.18 6 0.12
0.07 6 0.15

0.51 6 1.19** 20.02 6 0.32

were significantly associated with improvements in gait speed
(r = 0.20, P = 0.03) and chair stand times (r = 20.18, P = 0.05)
(Figure 2, A and B). No other significant correlations emerged
between serum magnesium supplementation and the other
0.17 6 0.67
0.17 6 0.63
0.17 6 0.70

0.33 6 1.78 21.77 6 0.55** 11.32 6 2.51 10.13 6 3.06 21.19 6 1.71** 11.79 6 2.44 11.46 6 2.20 20.33 6 1.79

0.03 6 0.18
physical performance tests considered. Changes in 24-h urinary

magnesium did not correlate with the results of the physical

performance tests.
Control group (n = 41)

A generalized linear model was used for a between-group comparison of the data obtained after 12 wk, adjusted for the baseline value of the corresponding test.
All values are means 6 SDs. *P , 0.05; **P , 0.001; ***P , 0.0001. RDA, Recommended Dietary Allowance; SPPB, Short Physical Performance Battery.
3.68 6 0.61 3.85 6 0.48
3.39 6 0.63 3.56 6 0.55
3.20 6 0.81 3.37 6 0.73

10.27 6 1.29 10.78 6 1.17

0.22 6 0.24*** 1.27 6 0.19 1.30 6 0.17

12 wk

Adverse events

Changes were calculated as the differences between data at baseline and at follow-up after 12 wk (within-group comparison) by using paired t tests.
No severe adverse events were reported in either group. Mild
Magnesium intake reaching RDA

adverse effects of magnesium supplementation included one


case of diarrhea and 2 of itching, which both regressed spon-

taneously after the treatment was suspended. Finally, no cases of
hypermagnesemia were observed in the group receiving sup-
3.68 6 0.55 3.50 6 0.88 20.18 6 0.95
0.58 6 0.72** 3.40 6 0.56 3.20 6 0.61 20.20 6 0.55 0.59 6 0.15*** 3.18 6 0.67 3.68 6 0.48 0.50 6 0.75
3.25 6 0.84 3.46 6 0.84 0.21 6 0.63

0.61 6 1.75

Magnesium oxide group (n = 29)

Values at baseline were obtained by means of a general linear models procedure including the interaction term in each model.

10.04 6 1.29 10.64 6 1.81

1.16 6 0.16 1.39 6 0.19

In the current study, we aimed to investigate the effect of

12 wk

magnesium supplementation on physical performance and muscle

strength in a sample of healthy, still physically active elderly
women. Our main findings were that magnesium supplementa-

tion significantly improved the total SPPB score and walking

speed and chair stand times. At the same time, the improvement
in total SPPB score in the treated group’s was twice as high as
change (12 wk)2

0.16 6 0.06**

0.90 6 0.35*
3.37 6 0.89 3.63 6 0.67 0.27 6 0.91 20.04 6 0.16

3.10 6 0.89 3.03 6 0.99 20.07 6 0.69 0.37 6 0.20

that in the control group and high enough to be clinically rele-

There were no statistically significant differences between the groups in any baseline test.

vant in terms of a meaningful change in physical performance

proposed for older people (29, 31).
Primary outcome in the treated and control groups, by baseline dietary magnesium intake1

Magnesium supplementation positively affected gait speed and

0.03 6 1.43

0.01 6 0.13

chair stand times. The improvement in gait speed was substantial:


the treated group had a mean improvement of w12 m/min vis-

Control group (n = 30)

à-vis the baseline. Like the improvement in total SPPB score, this
finding is important in clinical terms, because gait speed is the
0.83 6 1.24** 9.83 6 1.49 9.87 6 1.80

10.71 6 1.99 9.41 6 2.61 21.29 6 2.07** 12.07 6 2.72 12.40 6 3.08

0.21 6 0.29** 1.23 6 0.17 1.22 6 0.20

only item in the SPPB that can be used as a single variable for
12 wk

diagnosing sarcopenia (a condition involving a degenerative loss

of skeletal muscle mass), because it is an independent predictor
Magnesium intake below RDA

of adverse health events (23). Finally, magnesium supplemen-


tation strongly improved chair stand time, which appears to be

the most multidimensional item in the SPPB, including many
features essential to elderly people’s wellness, such as lower
0.13 6 0.54

0.13 6 0.85

limb strength, balance, and psychological aspects (32). On the

Magnesium oxide group (n = 24)

other hand, we found no differences in the balance tests, prob-

ably attributable to a ceiling effect. The role of a ceiling effect in
the SPPB test is further supported by the chair stands item: when
3.54 6 0.78 3.67 6 0.64
3.04 6 0.69 3.63 6 0.58
3.58 6 0.72 3.71 6 0.69

10.17 6 1.61 11.00 6 1.44

1.14 6 0.21 1.34 6 0.22

it was considered as a score, we found no significant difference

12 wk

between the groups, but when it was considered as a continuous

variable it showed the largest improvement among the SPPB

The effects of magnesium supplementation on the SPPB scores

were more evident in participants with a dietary magnesium
intake below the RDA, confirming previous reports that mag-
speed (m/s)

Chair stands
times (s)

nesium supplementation has beneficial effects on exercise per-




formance in magnesium-deficient individuals (16). Although we

only assessed dietary intake at baseline, any changes in diet were

Downloaded from

by guest
on 25 April 2018

deficiency, so the increase in their serum magnesium concen-

tration would reflect a positive effect in subjects with limited
magnesium reserves. This assumption is supported by the fact
that all our participants had normal serum magnesium concen-
trations at baseline, but about one-half of them had a low dietary
magnesium intake, 6% had serum magnesium concentrations
#0.75 mmol/L (39), and 5% had excreted magnesium concen-
trations ,2.5 nmol/d—the lower limit of normality at our lab-
oratory. Serum magnesium concentrations also proved clinically
more significant than urinary magnesium: we only found a sig-
nificant correlation between serum magnesium (but not urinary
magnesium) and changes in gait speed and chair stand times,
which suggests that that this simple test could be useful as
a first-line investigation when assessing physical performance in
older people.
In addition to the fact that dietary intake was investigated only
at baseline, the current study had some other limitations that need
to be mentioned. First, it was not a double-blind trial, although
only the research coordinator and main investigator knew
whether participants were taking magnesium supplements or not,
whereas the team members did not, and the outcomes were tested
by using machines or standardized procedures. It is impossible to
say how much of the differences found between the groups might
FIGURE 2. Associations between changes in serum magnesium and have been attributable to the intervention group’s expectations.
changes in Short Physical Performance Battery–determined gait speeds Moreover, this trial was conducted only in women; therefore,
(A) and chair stands times (B). Correlation analyses were performed by these findings cannot be generalized to men. On the other hand,
using Pearson’s correlations. D, change.
a strength of our work was its use of a global assessment of
physical performance and muscle strength, explored by using
reliable and reproducible methods. To the best of our knowl-
presumably limited given the short period of time covered by our edge, this was the first trial to include only elderly people and to
trial, so they would be unlikely to affect our findings. analyze the association of magnesium supplementation with
On the other hand, magnesium supplementation did not im- physical performance, considering both upper and lower limb
prove any of the secondary outcomes investigated, consistent strength items. Another strength was the comprehensive as-
with the findings of another study that explored magnesium sessment of magnesium status, which was investigated in terms
supplementation in middle-aged women (30). It is particularly of dietary intake, and serum and 24-h urinary magnesium con-
worth noting the apparent mismatch between the SPPB results centrations. Finally, body composition was assessed by dual-
and other outcomes relating to muscle strength. Considering the energy X-ray absorptiometry, which is the preferred method for
complexity of body-composition changes in the elderly, we
assessing body composition for research purposes and in the
hypothesize that the effect of magnesium supplementation may
clinical setting (23).
not have been apparent from the PT tests, which involve only
In conclusion, oral supplementation with 300 mg Mg in the
a few muscle units, and where significant differences emerge if
form of magnesium oxide for 12 wk had a significant positive
muscle function is associated with muscle mass changes. The
effect on physical performance, as assessed with the SPPB and
lack of any improvement in these tests is also reinforced by the
gait speed and chair stand times in healthy elderly women. These
participants’ training, which focused mainly on exercises de-
findings suggest a role for magnesium supplementation in pre-
signed to improve speed of contraction. The overall effect of
venting or delaying the age-related decline in physical perfor-
magnesium supplementation might therefore be more readily
mance, particularly in magnesium-deficient individuals. Further
detectable by using a detailed physical performance assessment,
research is needed to understand the influence of magnesium
such as the SPPB test. Further research is needed to clarify this
supplementation on physical performance in elderly people with
issue, however.
different magnesium concentrations.
As for the tertiary outcomes, magnesium supplementation for
12 wk led to a significant increase in both serum and urinary Our sincerest gratitude goes to the gym coordinator, Flavio Martinello, for
magnesium concentrations. Whereas oral magnesium supple- helping with recruitment. We thank the participants of the study, the nurses at
mentation is known to raise 24 h urinary magnesium (4), it is still the San Massimo Clinic (Padova Hospital, Laboratory Medicine), and Monica
a matter of debate whether it can increase serum magnesium and Elisabetta (Sanofi-Aventis), who helped us to find the free samples given
to the participants. Written permission was obtained from all persons named
concentrations in people with normal values. To date, only a trial
in this section.
with high doses of magnesium (600 mg) showed a significant
The authors’ responsibilities were as follows—NV, AC, and GS: designed
increase in serum magnesium in the treated group by comparison the trial; LB, SC, SP, GB, and FM: collected the data; EDT and EP: con-
with the control group (4, 30, 33–38). A possible reason why our ducted the analysis and interpreted the data; NV, LB, FB, and MDR: wrote
findings differed from those of other studies is that our women the manuscript; and EM and GS: took primary responsibility for the final
may have had an age-related subclinical form of magnesium content. None of the authors declared a conflict of interest.

Downloaded from

by guest
on 25 April 2018
REFERENCES 22. Heymsfield SB, Smith R, Aulet M, Bensen B, Lichtman S, Wang J,
1. Newhouse IJ, Finstad EW. The effects of magnesium supplementation Pierson RN Jr. Appendicular skeletal muscle mass: measurement by
on exercise performance. Clin J Sport Med 2000;10:195–200. dual-photon absorptiometry. Am J Clin Nutr 1990;52:214–8.
2. Barbagallo M, Belvedere M, Dominguez LJ. Magnesium homeostasis 23. Cruz-Jentoft AJ, Baeyens JP, Bauer JM, Boirie Y, Cederholm T, Landi
and aging. Magnes Res 2009;22:235–46. F, Martin FC, Michel JP, Rolland Y, Schneider SM, et al. Sarcopenia:
3. Wolf FI, Cittadini A. Chemistry and biochemistry of magnesium. Mol European consensus on definition and diagnosis: Report of the Euro-
Aspects Med 2003;24:3–9. pean Working Group on Sarcopenia in Older People. Age Ageing
4. Bohl CH, Volpe SL. Magnesium and exercise. Crit Rev Food Sci Nutr 2010;39:412–23.
2002;42:533–637. 24. Washburn RA, McAuley E, Katula J, Mihalko SL, Boileau RA. The
5. Lukaski HC. Magnesium, zinc, and chromium nutriture and physical physical activity scale for the elderly (PASE): evidence for validity.
activity. Am J Clin Nutr 2000;72(suppl):585S–93S. J Clin Epidemiol 1999;52:643–51.
6. Dominguez LJ, Barbagallo M, Lauretani F, Bandinelli S, Bos A, Corsi 25. Washburn RA, Smith KW, Jette AM, Janney CA. The Physical Activity
AM, Simonsick EM, Ferrucci L. Magnesium and muscle performance Scale for the Elderly (PASE): development and evaluation. J Clin
in older persons: the InCHIANTI study. Am J Clin Nutr 2006;84:419– Epidemiol 1993;46:153–62.
26. 26. Cameron ME, van Staveren WA. Manual of methodology for food
7. Rock E, Astier C, Lab C, Vignon X, Gueux E, Motta C, Rayssiguier Y. consumption studies. Oxford, United Kingdom: Oxford University
Dietary magnesium deficiency in rats enhances free radical production Press, 1988.
27. Nes M, van Staveren WA, Zajkas G, Inelmen EM, Moreiras O, Euronut
in skeletal muscle. J Nutr 1995;125:1205–10.
SENECA Investigators. Validity of the dietary history method in el-
8. Rude RK, Singer FR, Gruber HE. Skeletal and hormonal effects of
derly subjects. Eur J Clin Nutr 1991;45:S97–104.
magnesium deficiency. J Am Coll Nutr 2009;28:131–41.
28. Carnovale E, Miuccio FC. Tabelle di composizione degli alimenti.
9. Galan P, Preziosi P, Durlach V, Valeix P, Ribas L, Bouzid D, Favier A,
[Tables of food composition.] Rome, Italy: Ministry of Agriculture and
Hercberg S. Dietary magnesium intake in a French adult population.
National Institute of Nutrition, 1989 (in Italian).
Magnes Res 1997;10:321–8.
29. Kwon S, Perera S, Pahor M, Katula JA, King AC, Groessl EJ,
10. Ahlborg B, Ekelund LG, Nilsson CG. Effect of potassium-magnesium-
Studenski SA. What is a meaningful change in physical performance?
aspartate on the capacity for prolonged exercise in man. Acta Physiol
Findings from a clinical trial in older adults (the LIFE-P study). J Nutr
Scand 1968;74:238–45.
Health Aging 2009;13:538–44.
11. Ripari P, Pieralisi G, Giamberardino M, Vecchiet L. Effects of mag-
30. Moslehi N, Vafa M, Sarrafzadeh J, Rahimi-Foroushani A. Does mag-
nesium picolinate on some cardiorespiratory submaximal effort pa- nesium supplementation improve body composition and muscle
rameters. Magnes Res 1989;2:4. strength in middle-aged overweight women? A double-blind, placebo-
12. Brilla LR, Haley TF. Effect of magnesium supplementation on strength controlled, randomized clinical trial. Biol Trace Elem Res 2013;153:
training in humans. J Am Coll Nutr 1992;11:326–9. 111–8.
13. Weller E, Bachert P, Meinck HM, Friedmann B, Bartsch P, Mairbaurl 31. Perera S, Mody SH, Woodman RC, Studenski SA. Meaningful change
H. Lack of effect of oral Mg-supplementation on Mg in serum, blood and responsiveness in common physical performance measures in older
cells, and calf muscle. Med Sci Sports Exerc 1998;30:1584–91. adults. J Am Geriatr Soc 2006;54:743–9.
14. Terblanche S, Noakes TD, Dennis SC, Marais D, Eckert M. Failure of 32. McCarthy EK, Horvat MA, Holtsberg PA, Wisenbaker JM. Repeated
magnesium supplementation to influence marathon running perfor- chair stands as a measure of lower limb strength in sexagenarian
mance or recovery in magnesium-replete subjects. Int J Sport Nutr women. J Gerontol A Biol Sci Med Sci 2004;59:1207–12.
1992;2:154–64. 33. Hatzistavri LS, Sarafidis PA, Georgianos PI, Tziolas IM, Aroditis CP,
15. Finstad EW, Newhouse IJ, Lukaski HC, McAuliffe JE, Stewart CR. Zebekakis PE, Pikilidou MI, Lasaridis AN. Oral magnesium supple-
The effects of magnesium supplementation on exercise performance. mentation reduces ambulatory blood pressure in patients with mild
Med Sci Sports Exerc 2001;33:493–8. hypertension. Am J Hypertens 2009;22:1070–5.
16. Nielsen FH, Lukaski HC. Update on the relationship between mag- 34. Gontijo-Amaral C, Guimaraes EV, Camargos P. Oral magnesium
nesium and exercise. Magnes Res 2006;19:180–9. supplementation in children with cystic fibrosis improves clinical and
17. Tieland M, Brouwer-Brolsma EM, Nienaber-Rousseau C, van Loon LJ, functional variables: a double-blind, randomized, placebo-controlled
De Groot LC. Low vitamin D status is associated with reduced muscle crossover trial. Am J Clin Nutr 2012;96:50–6.
mass and impaired physical performance in frail elderly people. Eur J 35. Lukaski HC, Nielsen FH. Dietary magnesium depletion affects meta-
Clin Nutr 2013;67:1050–5. bolic responses during submaximal exercise in postmenopausal wom-
18. Schulz KF, Altman DG, Moher D. CONSORT 2010 statement: Updated en. J Nutr 2002;132:930–5.
guidelines for reporting parallel group randomised trials. J Pharmacol 36. Dimai HP, Porta S, Wirnsberger G, Lindschinger M, Pamperl I, Dobnig
Pharmacother 2010;1:100–7. H, Wilders-Truschnig M, Lau KH. Daily oral magnesium supple-
19. Institute of Medicine, Food and Nutrition Board. Dietary Reference mentation suppresses bone turnover in young adult males. J Clin En-
Intakes: calcium, phosphorus, magnesium, vitamin D and fluoride. docrinol Metab 1998;83:2742–8.
Washington, DC: National Academy Press, 1997. 37. Rodrı́guez-Moran M, Guerrero-Romero F. Oral magnesium supple-
20. Guralnik JM, Simonsick EM, Ferrucci L, Glynn RJ, Berkman LF, mentation improves insulin sensitivity and metabolic control in type 2
Blazer DG, Scherr PA, Wallace RB. A short physical performance diabetic subjects: a randomized double-blind controlled trial. Diabetes
battery assessing lower extremity function: association with self-reported Care 2003;26:1147–52.
disability and prediction of mortality and nursing home admission. 38. Song Y, He K, Levitan EB, Manson JE, Liu S. Effects of oral mag-
J Gerontol 1994;49:M85–94. nesium supplementation on glycaemic control in Type 2 diabetes:
21. Shepherd JA, Fan B, Lu Y, Wu XP, Wacker WK, Ergun DL, Levine a meta-analysis of randomized double-blind controlled trials. Diabet
MA. A multinational study to develop universal standardization of Med 2006;23:1050–6.
whole-body bone density and composition using GE Healthcare Lunar 39. Elin RJ. Assessment of magnesium status for diagnosis and therapy.
and Hologic DXA systems. J Bone Miner Res 2012;27:2208–16. Magnes Res 2010;23:S194–8.

Downloaded from

by guest
on 25 April 2018