Content uploaded by Hank Lukaski
Author content
All content in this area was uploaded by Hank Lukaski on Nov 07, 2018
Content may be subject to copyright.
The effects of magnesium supplementation
on exercise performance
ERIC W. FINSTAD, IAN J. NEWHOUSE, HENRY C. LUKASKI, JIM E. MCAULIFFE, and
CAMERON R. STEWART
School of Kinesiology, Lakehead University, Thunder Bay, Ontario, CANADA; and USDA Human Nutrition Research
Centre, Grand Forks, ND
ABSTRACT
FINSTAD, E. W., I. J. NEWHOUSE, H. C. LUKASKI, J. E. MCAULIFFE, and C. R. STEWART. The effects of magnesium
supplementation on exercise performance. Med. Sci. Sports Exerc., Vol. 33, No. 3, 2001, pp. 493–498. Purpose: To determine the
effects of magnesium (Mg
2⫹
) supplementation on performance and recovery in physically active women using the sensitive and
recently advanced measure of ionic Mg
2⫹
(iMg). Methods: Participants (N⫽121) were screened for [iMg] in plasma, with 44 (36.4%)
exhibiting [iMg] below the normal range of 0.53–0.67 mmol·L
-1
(4). Thirty-two subjects (21 ⫾3 yr) representing a broad range of
[iMg] (0.54 ⫾0.04 mmol·L
-1
) completed the main 14-wk study. At baseline, participants submitted to a resting blood pressure
measurement, and they completed both an anaerobic treadmill test and an incremental (aerobic) treadmill test. For the latter, values for
workload, oxygen uptake, and heart rate were obtained at both anaerobic threshold and maximal effort. Blood samples for iMg, total
serum Mg
2⫹
(TMg), erythrocyte Mg
2⫹
(EMg), Ca
2⫹
,K
⫹
,Na
⫹
, hemoglobin, hematocrit, lactate, and glucose were also collected
pretest, and 4, 10, 30 min, and 24 h posttest. Subjects received 212 mg·d
-1
Mg oxide or placebo in a double-blind fashion and were
retested after 4 wk. After a 6-wk washout period, the testing was repeated with a treatment crossover. Results: Ionic Mg
2⫹
increased
with Mg
2⫹
treatment versus placebo (P⬍0.05); however, performance and recovery indices were not significantly affected.
Conclusion: Four weeks of 212 mg·d
-1
Mg oxide supplementation improves resting [iMg] levels but not performance or recovery in
physically active women. Key Words: MAGNESIUM DEFICIENCY, IONIC MAGNESIUM, PHYSICALLY ACTIVE, WOMEN,
V
˙O
2
max, ANAEROBIC THRESHOLD, ANAEROBIC TREADMILL TEST, RECOVERY
Magnesium (Mg
2⫹
) is an essential mineral and co-
factor for over 400 enzymatic reactions with cen-
tral roles in the control of neuronal activity, car-
diac excitability, neuromuscular transmission, muscular
contraction, vasomotor tone, and blood pressure. The intra-
cellular functions of Mg
2⫹
are achieved through the forma-
tion of magnesium adenosine triphosphate (Mg
2⫹
-ATP)—a
substrate for a wide variety of enzymes that act to break-
down fatty acids, amino acids, and glucose during energy
metabolism (1). Magnesium plays a vital role in regulating
cell growth, reproduction, and membrane structure by reg-
ulating deoxyribonucleic acid and ribonucleic acid synthesis
and structure. Magnesium also serves as a Ca
2⫹
channel
blocker so that an inadequate amount of the mineral would
lead to hypertension and arrhythmias (1). The maintenance
of an adequate Mg
2⫹
status is therefore of utmost impor-
tance for optimal exercise performance and recovery.
Recommended intakes are 200 and 250 mg·d
-1
for women
and men, respectively, in Canada (16), and 280 and 350
mg·d
-1
, respectively, in the United States (14). As values
from 70 to 100% of the Recommended Dietary Allowance
(RDA) are considered adequate for an individual, cause for
concern results when intakes are less than 70% of the RDA
(17). Well over 50% of the normal population may have
marginal Mg
2⫹
deficiencies as they have Mg
2⫹
intakes
below that of the RDA (19). Lukaski (18) reports that
typical Western diets, which are high in protein and/or fat,
may not contain sufficient amounts of Mg
2⫹
. Also, due to
the increasing use of fertilizers (lacking Mg
2⫹
) and food
processing (removing Mg
2⫹
) in practice today, intakes have
declined from 500 mg·d
-1
to about 175–225 mg·d
-1
(2). In
addition, lower Mg
2⫹
intakes and, thus, marginal Mg
2⫹
deficiencies are more likely to be found in women rather
than men (7,17).
Concerning athletic populations, Seelig (24) notes that up
to half consume diets containing less than the RDA. Ath-
letes in particular tend to have increased needs for Mg
2⫹
most likely due to greater urinary and surface losses during
periods of exercise training (7,12,20,22). Because recom-
mended intakes of the mineral are based on nonathletic
populations, and because the literature regarding the actual
Mg
2⫹
status of athletes is limited, one can still question
whether the observed dietary intakes are sufficient for the
amount of energy that the athletes are expending and the
potential increased Mg
2⫹
losses.
Research has been equivocal in regard to Mg
2⫹
supple-
mentation’s effect on exercise performance. Some studies
have been supportive (9,10,13,15,21), whereas others
(23,25,27–29) found no benefit to exercise performance.
Although it may be assumed that a suboptimal intake of
Mg
2⫹
could result in physiological impairments, research
0195-9131/01/3303-0493/$3.00/0
MEDICINE & SCIENCE IN SPORTS & EXERCISE
®
Copyright © 2001 by the American College of Sports Medicine
Submitted for publication February 2000.
Accepted for publication May 2000.
493
has relied on insensitive indicators of Mg
2⫹
status if Mg
2⫹
status was measured at all.
It is therefore not clear whether Mg
2⫹
supplementation,
beyond the maintenance of an adequate dietary intake of the
mineral, is effective in enhancing performance and recovery
from exercise. In addition, little research has concentrated
on physically active women who may be at the highest risk
for Mg
2⫹
deficiency. Thus, the purpose of this investigation
was to examine the effects of Mg
2⫹
supplementation on
Mg
2⫹
levels as well as the effects of Mg
2⫹
supplementation
on performance and recovery in physically active women.
This would be achieved by utilizing the highly specific,
sensitive, and recently advanced measure of ionic Mg
2⫹
(iMg) assay, and by employing a randomized, double-blind,
placebo-controlled crossover experimental design. It was
hypothesized that Mg
2⫹
supplementation would increase
Mg
2⫹
levels as well as performance and recovery from
exercise.
METHODS
Participants. Once written informed consent was ob-
tained (Ethics Advisory Committee, Lakehead University),
121 apparently healthy, physically active women between
the ages of 17 and 43 residing in the Thunder Bay, Ontario,
Canada, region were screened for participation based on
iMg status. Those who qualified to continue the study were
20 marginally Mg
2⫹
-deficient subjects and 20 who pos-
sessed [iMg] levels in the upper range of normal. A power
test utilizing previous test data on V
˙O
2
max values (9) indi-
cated the sample size of 40 should be sufficient to reach a
power of 0.80 in detecting an effect size of 3 mL·kg
-1
·min
-1
.
The normal range for [iMg] is 0.53–0.67 mmol·L
-1
(4).
Although the intent of screening based on iMg status was to
permit later statistical analysis with initial [iMg] values
being an independent variable, the lability of [iMg] made
this proposition problematic. Over the course of 3 wk from
the time of screening to the date of the first testing period,
many subjects changed their iMg classification from low to
normal. As well, the correlation of iMg values between the
two time periods when no treatment had been initiated was
low (see Results). Selection criteria for subjects included: A.
no clinical history of proven or suspected hypersensitivity to
Mg
2⫹
supplements; B. the expectation and willingness to
maintain regular physical activity (i.e., performing at least
three workouts at or above 75% maximum intensity, for at
least 2 h total, per week) throughout the study; and C. the
expectation and willingness to maintain regular dietary in-
takes throughout the study. Exclusion criteria for subjects
included: A. not performing the incremental treadmill test at
all four test sessions, B. not ingesting at least 75% of pills
given per treatment period, C. being ill or injured (enough to
prevent exercising) for more than 1 wk during the treatment
periods, and D. not maintaining a training load equal to ⫾
25% of their regular load. Enrolment of participants con-
tinued until the requisite number of subjects had been
identified.
Measurements. Selected subjects underwent three day
dietary analysis and anthropometric measurements (height,
weight, and sum of five skinfolds). Physiological testing
included: A. resting blood pressure; B. incremental tread-
mill test with measurement of workload, heart rate, and
expired gases—for assessment of V
˙O
2
max, anaerobic
threshold, maximal workload, and submaximal running ef-
ficiency; C. blood sampling pretest, and 4, 10, 30 min, and
24 h posttest, for assessment of [iMg], [TMg], [iCa], [plas-
ma lactate], [plasma glucose], [plasma K
⫹
], [plasma Na
⫹
],
[hemoglobin], hematocrit; D. anaerobic treadmill test; and
E. subjective and objective measures of overtraining re-
corded daily in a training diary (for control purposes).
Four testing periods (T1, T2, T3, and T4) were scheduled.
During the weeks of T1 and T3, subjects filled out three
consecutive days (including one weekend day) of self-report
dietary records. Dietary intakes were assessed using com-
puterized diet analysis software (Diet Analysis Plus
TM
,
West Publishing Co., St. Paul, MN). Subjects were asked to
refrain from unaccustomed strenuous exercise for 48 h (and
any strenuous exercise 24 h) before exercise testing. They
were also asked to refrain from alcohol consumption 24 h
(and caffeine consumption 6 h) before testing. They were
advised to consume a light carbohydrate meal 2–3 h before
reporting to the exercise tests which took place between
1:00 and 9:00 p.m. Anthropometric measurements were
performed by a Certified Fitness Appraiser according to
national standards (11). Harpenden calipers were used to
measure skinfolds that were taken from the triceps, biceps,
subscapular, iliac crest, and medial calf regions. Percent
body fat was predicted using the Durnin-Womersley method
(11).
Resting blood pressure (performed by a Certified Fitness
Appraiser using an AMG Med. Professional Series sphyg-
momanometer) and resting heart rate (using a Polar Vantage
XL® heart rate monitor from Polar CIC Inc., Kempele,
Finland) were obtained after the subject rested in the sitting
position for 10 min. Blood was collected in 7-mL green-
topped Vacutainer® tubes (lithium-heparin added) by ante-
cubital venipuncture. The tourniquet was applied gently and
released before the actual blood draw.
The multi-test Stat Profile
TM
Ultra Analyzer® model
11–3C (Nova Biomedical Canada Ltd., Mississauga, On-
tario) was utilized for the immediate analyses of [iMg],
[iCa], [Na
2⫹
], [K
⫹
], [hemoglobin], hematocrit, [glucose],
and [lactate] from 25
g of whole blood. The instrument
was housed in the same laboratory as the exercise testing,
and the analyses were performed by the same technician
throughout the study. The precision of the instrument for
[iMg] was estimated at ⫾0.03 mmol·L
-1
, as determined by
repeat testing on the same sample of blood, both within and
between days, on an individual not involved in the study. All
analyses were within their respective reference ranges as
determined by NOVA quality control substances.
Before the incremental treadmill test, the subjects
warmed up on the treadmill (Quinton Instruments, Seattle,
WA) for 5 min. The initial treadmill speed was 2.22 m·s
-1
and during the test was increased by 0.22 m·s
-1
every
494
Official Journal of the American College of Sports Medicine http://www.acsm-msse.org
minute. The treadmill grade remained horizontal until the
subject had completed two workloads past the workload at
which the subject’s respiratory exchange ratio (expired
CO
2
/inspired O
2
) passed a value of 1.0. At that time, the
speed no longer increased but the grade increased by 2%
each min. Subjects continued until exhaustion. Expired
gases were sampled and analyzed by a SensorMedics V
max
System® metabolic (Yorba Linda, CA) with V
˙O
2
max and
V
E
peak determined as the highest 30-s mean. Heart rates
were monitored by a telemetric heart rate monitor and
recorded 30 s into each workload. Anaerobic thresholds
were assessed via examination of expired gas values ac-
cording to the method described by Beaver et al. (6). Two
investigators blinded to the study, and familiar with anaer-
obic threshold estimations, independently analyzed the V
E
/
VCO
2
curves to pick out the lowest point. These researchers
then met to compare estimates. Discrepancies (more than
one min differences) were resolved by examining other
expired gas indices of anaerobic threshold. Efficiency of
running at a submaximal workload was assessed by noting
the workload during the T1 test that corresponds to 65% of
V
˙O
2
max. At T2, T3, and T4, the oxygen consumption was
noted at this same workload.
The anaerobic treadmill test was performed 48 h after the
incremental treadmill test, and according to the protocol
described by Bouchard et al. (8). In brief, the test required
the subjects to exert a maximal effort on the treadmill with
the speed set at 3.31 m·s
-1
and the grade set at 20%. Time
to exhaustion (approximately 30–90 s) was the only vari-
able measured.
Encouragement was given equally to all subjects for all
treadmill tests. The protocols described above remained
identical for each of the four testing periods. The timing in
terms of day of week and time of day were consistent for
each subject with a few exceptions to accommodate sub-
jects’ schedules.
The screening of the 121 subjects for [iMg] was con-
ducted over 3 d. Three weeks later, the first testing session
(T1) began. Each of the four testing sessions (T1, T2, T3,
and T4) involved the following: Day 1 consisted of the
distribution of dietary records (T1 and T3 only), and sub-
mission to anthropometric tests, resting heart rate and blood
pressure measurement, and an incremental treadmill test
with pretest, and 4, 10, and 30 min posttest blood with-
drawal. Day 2 involved only the 24-h posttest blood with-
drawal. Day 3 consisted of the anaerobic treadmill test and
the dispensing of pills (Mg
2⫹
supplement or placebo) (T1
and T3 only).
In a double-blind manner, subjects were randomly as-
signed to begin treatment with either 212 mg·d
-1
(two pills
of 106 mg elemental Mg
2⫹
) Mg oxide (C. E. Jamieson and
Company Ltd., Windsor, Ontario) or matching placebo. The
placebo contained 222 mg dicalcium phosphate, 12 mg
purified stearic acid, 6 mg coscarmellose sodium, 4 mg
silicon dioxide, and 276 mg microcrystalline cellulose. Both
the Mg
2⫹
supplement and the placebo were in tablet form,
and were identical in appearance, consistency, and taste.
Subjects were instructed to consume the contents of one
tablet twice per day (one just before breakfast and one just
before dinner). The treatment lasted 4 wk (28 ⫾3 d), which
was followed by a 6-wk (42 ⫾3 d) washout period. Treat-
ments were then reversed for the final 4 wk.
A randomized, double-blind, crossover (Mg oxide vs.
placebo) design was employed. Testing (identical to that
noted above) occurred every 4 wk (at T1, T2, T3, and T4).
Subjects were divided into two treatment groups, one for
which the order of pill administration was Mg
2⫹
supplement
first and placebo second (group M/P), and the other, vice
versa (group P/M).
Statistical analysis. Correlational coefficients were
used to establish relationships between the baseline mea-
sures of dietary intakes, hematological assays, and perfor-
mance variables. The dietary intakes and training diaries
were examined so that any discrepancies in routine would be
taken into account. To test whether a carryover effect was
present (i.e., if the washout period was ineffective) an in-
dependent t-test was performed on the [iMg] change scores
from pre to posttreatment (5). Once the carry-over effect
was disclaimed (see Results), the data were pooled to test
for differences between treatment groups. The differences
between treatments were measured using Student’s t-test for
paired samples on change scores from baseline values to
treated values. To investigate whether training state affects
the efficacy of Mg
2⫹
supplementation, ANOVAs were run.
Training state was classified as above or below the T1 mean
using V
˙O
2
max values and again using time to exhaustion on
the anaerobic treadmill test. The accepted level of signifi-
cance was P⬍0.05 for all statistical tests.
RESULTS
The characteristics of the 121 subjects screened for the
study are listed in Table 1. Of the 121 originally screened for
this study, 44 (or 36.4%) were marginally Mg
2⫹
-deficient
according the criteria set by Altura et al. (4), i.e., they
exhibited an [iMg] of less than 0.53 mmol·L
-1
. Most sub-
jects maintained their training regimens throughout the
study, although there was some seasonal fluctuation, espe-
cially as several did not exercise consistently during the
washout period, which took place over the Christmas sea-
son. However, no differences in training, lifestyle, or dietary
intake were apparent between treatment groups. Subjects
participated in sporting activities, which were aerobic (most
often running, cross-country skiing, and swimming) and/or
anaerobic (most often weight training, soccer, basketball,
and volleyball). All participants possessed regular menstrual
patterns and none reported gastrointestinal distress through-
out the study.
TABLE 1. Characteristics of screened subjects (
N
⫽121 subjects).
Parameter Means ⴞSD Range
Age (yr) 21.5 ⫾4.2 17–43
Height (m) 165.8 ⫾6.2 151.5–178.0
Weight (kg) 63.2 ⫾8.2 45.0–87.0
Resting [iMg] level (mmol䡠L
⫺1
)0.54 ⫾0.04 0.46–0.69
Systolic blood pressure (mm Hg) 109.9 ⫾8.8 94–130
Diastolic blood pressure (mm Hg) 66.8 ⫾7.9 50–90
MAGNESIUM TREATMENT ON EXERCISE PERFORMANCE Medicine & Science in Sports & Exercise姞
495
Before completing the study, seven subjects dropped out
from group M/P (N⫽13 as a result), and one dropped out
from group P/M (N⫽19 as a result). Of these eight
individuals, two became injured, three became ill, and three
experienced scheduling difficulties; none were included in
the analysis.
Table 2 lists the characteristics of the 32 subjects com-
pleting the entire study as well as the characteristics of both
treatment groups, M/P and P/M, with independent t-tests
performed between groups. Because there were no signifi-
cant differences between treatment groups on baseline mea-
sures, further analyses were justified. The mean dietary
Mg
2⫹
intakes were 320 ⫾123 mg·d
-1
at T1 and 333 ⫾127
mg·d
-1
at T2. Each of the two sets of dietary analyses
revealed that six (or 20%) had intakes less than the Canadian
Recommended nutrient intake (RNI) amount of 200 mg of
Mg
2⫹
, and the same number had less than the 70% RDA
amount. There was no correlation between Mg
2⫹
intake and
[iMg] (r
(32)
⫽⫺0.27) at T1.
In studies with crossover designs such as in this study, it
is necessary to test for a carry-over effect, i.e., to see if the
supplemental Mg
2⫹
(or the placebo) still carried any effect
from the first treatment period to the second treatment
period. Armitage and Hills (5) also refer to this effect as the
“treatment by period” interaction, where “treatment” is the
factor representing the Mg
2⫹
supplementation or placebo,
and “period” is the factor representing the two periods of
treatment. An independent t-test was performed between
treatment groups on [iMg] change scores from pre to post-
treatment (t
(30)
⫽⫺0.05, P⬎0.05). Because the test did not
reveal a treatment by period interaction, the carry-over ef-
fect was dismissed; thus, data from the two treatment peri-
ods were pooled (or combined) to form two categories—
Mg
2⫹
-treated (M), and placebo-treated (P).
The lability of [iMg], as was noted in Methods, con-
founded its use as an independent variable. To highlight the
lability of [iMg], in the 3-wk interval between the initial
screening of potential subjects and T1, resting [iMg] signif-
icantly increased from 0.532 ⫾0.046 mmol·L
-1
to 0.567 ⫾
0.032 mmol·t
-1
,t
(31)
⫽4.70; P⬍0.05 (N⫽32). Whereas
50% would have been classified as marginally deficient
using the screening results, only 15.6% were marginally
deficient at T1. The correlation for resting [iMg] levels
between screening and T1 was r ⫽0.45. Lability can also be
highlighted by noting correlations between resting [iMg]
levels from pretreadmill test to 24 h posttest and these were
r⫽0.55 (T1), r ⫽0.35 (T2), r ⫽0.80 (T3), and r ⫽0.59
(T4).
Despite lability concerns, pooled data revealed that the
increase in resting [iMg] was significantly more for M
(⫹0.044 mmol·L
-1
)thanP(⫹0.028 mmol·L
-1
), t
(31)
⫽2.28;
P⬍0.05 (see also Table 3). Performance and recovery
index values did not differ significantly between treatments.
However, a trend was indicated for V
E
peak values as they
increased slightly more for M (⫹2.25 L·min
-1
)vsP(⫺2.01
L·min
-1
)(t
(25)
⫽1.75; P⫽0.09).
Pooled data revealed no differences in Mg
2⫹
supplemen-
tation’s affect between those with the higher and lower
V
˙O
2
max values (P⫽0.50) nor when using higher vs lower
time to exhaustion scores on the anaerobic treadmill test
(P⫽0.60).
DISCUSSION
Results indicated that 4 wk of treatment with 212 mg·d
-1
Mg oxide was successful in raising resting [iMg] levels.
Previous research has discovered slight but insignificant
increases in [TMg] and [EMg] (after three weeks of sup-
plementation with 387 mg·d
-1
Mg pidolate) (26). It may be
that the [iMg] measure is more sensitive than [TMg] or
[EMg] as previously claimed (2,3). Therefore, the use of this
assay to assess the effects of Mg
2⫹
treatment on perfor-
mance and recovery may be advantageous over other he-
matological measures of Mg
2⫹
.
The significant increase in [iMg] levels from screening to
T1 may be due to the subjects’ increased awareness and
interest about their nutritional habits from their involvement
in such a study. Natural regression toward the mean may
have also led to this increase. Because there were mostly
weak correlations between screening and T1 [iMg] values and
between pretest and 24 h posttest resting [iMg] values, it seems
that these levels were fairly labile within individuals. The
instrument itself does not appear to account for this as the
precision (coefficient of variation) of the ion-selective elec-
trode for Mg
2⫹
has been reported to be excellent, i.e., less than
6% for control samples (4), and indeed quality control samples
TABLE 2. Baseline characteristic of subjects completing entire study.
Parameter All Subjects Group M/P Group P/M
t
-Test Result
Age (yr) 21.2 ⫾3.1 20.8 ⫾1.8 21.4 ⫾3.8
t
(30) ⫽⫺0.46
Height (cm) 165.5 ⫾5.6 166.3 ⫾7.2 165.0 ⫾4.2
t
(30) ⫽0.63
Weight (kg) 61.5 ⫾5.3 61.9 ⫾7.1 61.2 ⫾3.8
t
(30) ⫽0.34
Predicted percent body fat (%) 22.5 ⫾4.0 23.1 ⫾4.4 22.1 ⫾3.8
t
(30) ⫽0.70
Resting [iMg] level (mmol䡠L
⫺1
)0.57 ⫾0.03 0.56 ⫾0.04 0.57 ⫾0.03
t
(30) ⫽⫺1.04
Mg
2⫹
intake (mg䡠d
⫺1
)320.4 ⫾123.2 305.8 ⫾133.5 325.6 ⫾115.0
t
(30) ⫽⫺0.45
V
˙O
2
max (mL䡠min
⫺1
䡠kg
⫺1
)49.8 ⫾6.3 49.4 ⫾6.8 50.2 ⫾6.1
t
(30) ⫽⫺0.34
V
E
peak (L䡠min) 109.2 ⫾12.3 109.3 ⫾11.1 109.2 ⫾13.3
t
(30) ⫽0.04
Maximum heart rate (beats per min) 195.1 ⫾9.5 200.0 ⫾7.2 193.9 ⫾9.5
t
(30) ⫽1.96
V
˙O
2
at an aerobic threshold (mL䡠min
⫺1
䡠kg
⫺1
)37.4 ⫾4.6 38.3 ⫾4.9 36.7 ⫾4.4
t
(30) ⫽0.94
Anaerobic treadmill test result(s) 41.7 ⫾14.3 41.0 ⫾16.8 42.2 ⫾12.8
t
(26) ⫽0.24
Systolic blood pressure (mm Hg) 114.3 ⫾8.9 116.3 ⫾9.9 113.0 ⫾8.1
t
(30) ⫽1.04
Diastolic blood pressure (mm Hg) 69.4 ⫾9.5 70.2 ⫾12.4 69.0 ⫾7.3
t
(30) ⫽0.35
Group M/P, magnesium—1
st
, placebo—2
nd
(
N
⫽13); Group P/M, placebo—1
st
, magnesium—2
nd
(
N
⫽19). Total of 32 subjects or 30 for Mg
2⫹
intake; means ⫾SD. Independent
t
-tests performed on treatment groups; *
P
⬍0.05.
496
Official Journal of the American College of Sports Medicine http://www.acsm-msse.org
and repeated sampling on the same blood yielded consistent
values in the present study. Physiological effects could thus be
deduced to be the cause of the variability, although at this point
the source of the physiological perturbation is unknown. Some
control was in place regarding previous exercise and diet in-
cluding alcohol and caffeine consumption. The variability of
[iMg] confounded clarification of the role Mg
2⫹
status plays in
the efficacy of Mg supplementation. The variability of [iMg]
within individuals over time must be researched in more detail
as the assessment, classification and diagnosis of Mg
2⫹
status
based on these values are put in jeopardy.
The mean dietary Mg
2⫹
intakes of 320 ⫾123 mg·d
-1
(for
T1) and 333 ⫾127 mg·d
-1
(for T2) are well above both the
RNI of 200 mg·d
-1
and the RDA of 280 mg·d
-1
. Altura (2)
claimed that female subjects (without regard to activity
level) ingest approximately 175–225 mg·d
-1
Mg
2⫹
(60–
80% of the RDA), whereas Lukaski (18) noted that female
athletes take in about 168–182 mg·d
-1
of the mineral (60–
65% of the RDA). It seems that participants in this study
possess Mg
2⫹
intakes that are higher than normal. Apart
from being regular exercisers between the ages of 17 and 29,
the subjects appeared to be quite diverse in regard to dietary
habits. Some subjects gave conscientious effort to obtaining
a healthy diet, whereas others tended toward highly refined
convenience foods that often lack Mg
2⫹
in adequate
amounts. In addition, because there was no correlation be-
tween Mg
2⫹
intakes and [iMg] levels (r
(32)
⫽⫺0.27), it
may be that there is a wide variability in the way in which
Mg
2⫹
is metabolized. This may have meant that the sup-
plementation would have had less effect on some
participants.
There were no significant effects of Mg
2⫹
supplementa-
tion on performance during (or recovery from) aerobic or
anaerobic exercise. Previous studies have discovered bene-
fits of Mg
2⫹
treatment on O
2
(10, 22) consumption at
submaximal workloads, total workload (26), and time to
exhaustion (9). Magnesium treatment has also been shown
to improve submaximal performance by decreasing sub-
maximal V
˙O
2
,V
E
, and HR (9, 21). Yet all of these studies
involved higher dosages of Mg
2⫹
supplementation and ei-
ther used subjects with lower mean Mg
2⫹
intakes or did not
report intakes at all. Moreover, only Vecchiet and col-
leagues (26) incorporated the highly controlled crossover
experimental design as utilized in this study. These choices
may have contributed extensively to reporting significant
effects. In addition, it is difficult to speculate on the possi-
bility of Mg
2⫹
deficiencies in the research of Brilla and
Gunter (9) as well as that of Ripari and coworkers (21)
because Mg
2⫹
status was not assessed. The results of the
present study are more in line with those of Ruddel et al.
(23), Weight et al. (28), and Weller et al. (29). Weight and
coworkers utilized a crossover design and a lower dose of
Mg
2⫹
(116 mg·d
-1
) and found no significant effects on
performance in subjects with high Mg
2⫹
intakes (372 ⫾122
mg·d
-1
). The only study to specifically screen for a low
[TMg] was that of Weller et al., and they did not find a
benefit of Mg supplementation on exercise performance in
subjects whose initial [TMg] values were in the low-normal
range.
On the other hand, it is possible that the present study is
lacking in certain areas that may have been critical for
obtaining positive effects of Mg
2⫹
supplementation. The
relatively high mean Mg
2⫹
intake of the participants and the
relatively low dosage of Mg
2⫹
supplementation may have
prevented the opportunity for significant increases in ab-
sorption of the mineral in the body. Although this study built
on previous research in the area by utilizing the sensitive
[iMg] measure to assess Mg
2⫹
status, the discovery that
[iMg] appears to be rather labile compromises its usefulness
(at least until further research can elucidate all of the phys-
iological factors contributing to the variability). As well,
because reference ranges for [iMg] levels have not yet been
well established, it is difficult to confirm whether some
subjects were initially Mg
2⫹
-deficient or not.
There appears to be nothing to suggest that initial training
state (specifically V
˙O
2max
and anaerobic treadmill test
TABLE 3. Physiological testing parameters by treatment.
Physiological Variable Treatment
Raw Scores Change
t
-Test ResultBaseline Treated Actual %
Resting [iMg] (mmol䡠L
⫺1
)M 0.563 ⫾0.045 0.607 ⫾0.037 ⫹0.044 ⫹7.82
P 0.563 ⫾0.038 0.591 ⫾0.043 ⫹0.028 ⫹4.97
t
(31) ⫽2.28*
V
˙O
2
max (mL䡠min
⫺1
䡠kg
⫺1
)M 49.66 ⫾6.54 50.70 ⫾6.65 ⫹1.04 ⫹2.09
P 49.77 ⫾5.96 50.20 ⫾6.69 ⫹0.43 ⫹0.86
t
(26) ⫽0.95
V
E
peak (L䡠min
⫺1
)M 107.21 ⫾13.59 109.46 ⫾14.47 ⫹2.25 ⫹2.10
P 107.54 ⫾12.64 105.53 ⫾13.95 ⫺2.01 ⫺1.87
t
(25) ⫽1.75
Maximal workload (% grade) M 3.25 ⫾2.63 3.13 ⫾2.43 ⫺0.12 ⫺3.69
P 3.66 ⫾1.86 3.59 ⫾2.35 ⫺0.07 ⫺1.91
t
(28) ⫽⫺0.14
Maximal heart rate (beats per min) M 197.00 ⫾8.35 195.81 ⫾7.84 ⫺1.19 ⫺0.60
P 195.34 ⫾9.54 195.52 ⫾6.99 ⫹0.18 ⫹0.09
t
(28) ⫽⫺0.66
V
˙O
2
at AT (mL䡠min
⫺1
䡠kg
⫺1
)M 38.31 ⫾3.95 39.12 ⫾4.30 ⫹0.81 ⫹2.11
P 38.16 ⫾5.40 38.75 ⫾4.20 ⫹0.59 ⫹1.55
t
(30) ⫽0.22
Anaerobic test (s) M 36.54 ⫾14.74 37.18 ⫾13.50 ⫹0.64 ⫹1.75
P 38.47 ⫾11.87 37.80 ⫾13.72 ⫺0.67 ⫺1.74
t
(27) ⫽1.06
Systolic BP (mm Hg) M 115.09 ⫾9.45 110.94 ⫾7.72 ⫺4.15 ⫺3.61
P 114.31 ⫾8.53 112.69 ⫾8.91 ⫺1.62 ⫺1.42
t
(31) ⫽⫺1.26
Diastolic BP (mm Hg) M 68.94 ⫾9.03 68.00 ⫾7.73 ⫺0.94 ⫺1.36
P 68.84 ⫾6.91 68.59 ⫾7.80 ⫺0.25 ⫺0.36
t
(31) ⫽⫺0.36
Pooled data based on treatment: M, magnesium; P, placebo. Baseline and Treated, tests performed before and after treatment, respectively (means ⫾SD). Student’s
t
-tests performed
on change scores; *
P
⬍0.05.
MAGNESIUM TREATMENT ON EXERCISE PERFORMANCE Medicine & Science in Sports & Exercise姞
497
performance) affects the response to Mg
2⫹
supplementation
in this group of subjects. A review of the literature may have
suggested otherwise because those studies revealing no ef-
fects with Mg
2⫹
supplementation (23,25,27,29) all incorpo-
rated trained subjects, whereas those finding a positive ef-
fect (9,10,13,15,21,26) predominantly used untrained
subjects. The homogeneity of this study’s subject pool may
not have permitted a clear distinction between the higher
and lower training status groupings.
This study concluded that 4 wk of 212 mg·d
-1
Mg oxide
supplementation significantly improved resting [iMg] levels
but not performance or recovery in a group of physically
active females.
This study was supported by the Gatorade Sports Science Insti-
tute (Barrington, IL), Nova Biomedical Canada Limited (Mississauga,
ON), the United States Department of Agriculture Human Nutrition
Research Center (Grand Forks, ND), and C. E. Jamieson and Com-
pany Limited (Windsor, ON). Note that the results of this study do
not constitute endorsement of the products used in this study by the
authors or ACSM.
Address for correspondence: Dr. Ian Newhouse, Lakehead
University, Thunder Bay, Ontario, P7B 5E1 Canada; E-mail: ian.
newhouse@lakeheadu.ca.
REFERENCES
1. ALTURA, B. M. Basic biochemistry and physiology of magnesium: A
brief review. Magnesium and Trace Elements 10:167–171, 1991–92.
2. ALTURA, B. M. Importance of magnesium in physiology and med-
icine and the need for ion sensitive electrodes. Scand. J. Clin. Lab.
Invest. 54(Suppl. 217):83–89, 1994.
3. ALTURA, B. T., J. L. BURACK,R.Q.CRACCO, et al. Clinical studies
with the NOVA ISE for iMg
2⫹
.Scand. J. Clin. Lab. Invest.
54(Suppl. 217):53–68, 1994.
4. ALTURA, B. T., T. L. SHIREY,C.C.YOUNG, et al. Characterization
of a new ion selective electrode for ionized magnesium in whole
blood, plasma, serum, and aqueous samples. Scand. J. Clin. Lab.
Invest. 54(Suppl. 217):21–36, 1994.
5. ARMITAGE, P., and M. HILLS. The two-period crossover trial. The
Statistician 31:119–131, 1982.
6. BEAVER, W. L., K. WASSERMAN, and B. J. WHIPP. A new method for
detecting anaerobic threshold by gas exchange. J. Appl. Physiol.
60:2020–2027, 1986.
7. BOHMER, D. Magnesium deficit: Relation to physical performance.
In: Magnesium and Physical Activity, L. Vecchiet (Ed.). New
York: Parthenon, 1995, pp. 149–156.
8. BOUCHARD, C., A. W. TAYLOR, and S. DULAC. Testing maximal
anaerobic power and capacity. In: Physiological Testing of the
Elite Athlete, J. D. MacDougall, H. A. Wenger, and H. J. Green
(Eds.). Canada: Canadian Association of Sport Sciences, Mutual
Press, 1982, pp. 61–74.
9. BRILLA, L. R., and K. B. GUNTER. Effect of magnesium supple-
mentation on exercise time to exhaustion. Med. Exerc. Nutr.
Health 4:230–233, 1995.
10. BRILLA, L. R., and T. F. HALEY. Effect of magnesium supplemen-
tation on strength training in humans. J. Am. Coll. Nutr. 11:326–
329, 1992.
11. CANADIAN SOCIETY FOR EXERCISE PHYSIOLOGY.Certified Fitness
Appraisal Resource Manual. Gloucester, ON: Canadian Society
for Exercise Physiology, 1995, pp. 111–113.
12. DEUSTER, P., A., DOLEV,S.B.KYLE,R.A.ANDERSON, and E. B.
SCHOOMAKER. Magnesium homeostasis during high-intensity an-
aerobic exercise in men. J. Appl. Physiol. 62:545–550, 1987.
13. DRAGANI, L., M. A. GIAMBERARDINO, and L. VECCHIET. Effects of
magnesium administration on muscle damage from physical ex-
ercise. In: Magnesium and Physical Activity, L. Vecchiet (Ed.).
New York: Parthenon, 1995, pp. 253–260.
14. NATIONAL RESEARCH COUNCIL (U.S.)/SUBCOMMITTEE ON THE TENTH
EDITION OF THE RDAS.Recommended Dietary Allowances, 10th
Ed. Washington, DC: National Academy Press, 1989.
15. GOLF, S. W., S. BENDER, and J. GRUTTNER. On the significance of
magnesium in extreme physical stress. Cardiovasc. Drugs Ther.
12(Suppl. 2):197–202, 1998.
16. HEALTH AND WELFARE CANADA.Nutrition recommendations: Re-
port of the scientific review committee. Ottawa, ON: Minister of
Supply and Service Canada, 1990, pp. 136–140.
17. LUKASKI, H. C. Micronutrients (magnesium, zinc, and copper): are
mineral supplements needed for athletes? Int. J. Sport Nutr.
5:S74–83, 1995.
18. LUKASKI, H. C. Prevention and treatment of magnesium deficiency
in athletes. In: Magnesium and Physical Activity, L. Vecchiet
(Ed.). New York: Parthenon, 1995, pp. 211–226.
19. MCDONALD, R., and C. L. Keen. Iron, zinc and magnesium nutri-
tion and athletic performance. Sports Med. 5:171–184, 1988.
20. RESINA, A., L. GATTESCHI,M.G.RUBENNI, et al. Changes in serum
and erythrocyte magnesium after training and physical exercise.
In: Magnesium and Physical Activity, L. Vecchiet (Ed.). New
York: Parthenon, 1995, pp. 199–209.
21. RIPARI, P., G. PIERALISI,M.A.GIAMBERARDINO,A.RESINA, and
L. VECCHIET. Effects of magnesium pidolate on some cardiore-
spiratory submaximal effort parameters. Magnesium Res. 2:70,
1989.
22. ROSE, L. I., D. R. CARROLL,S.L.LOWE,E.W.PETERSON, and K. H.
COOPER. Serum electrolyte changes after marathon running.
J. Appl. Physiol. 29:449–451, 1970.
23. RUDDEL, H., C. WERNER, and H. ISING. Impact of magnesium
supplementation on performance data in young swimmers. Mag-
nes. Res. 3:103–107, 1990.
24. SEELIG, M. S. Consequences of magnesium deficiency on the
enhancement of stress reactions: preventive and therapeu-
tic implications (a review). J. Am. Coll. Nutr. 13:429–446, 1994.
25. TERBLANCHE, S., NOAKES, T. D., DENNIS, S. C., D. W. MARAIS, and
M. ECKERT. Failure of magnesium supplementation to influence
marathon running performance or recovery in magnesium replete
subjects. J. Sports Nutr. 2:154–164, 1992.
26. VECCHIET, L., G. PIERALISI,M.D’OVIDIO, et al. Effects of magne-
sium supplementation on maximal and submaximal effort. In:
Magnesium and Physical Activity, L. Vecchiet (Ed.). New York:
Parthenon, 1995, pp. 227–237.
27. WEIGHT, L. M., K. H. MYBURGH, and T. D. NOAKES. Vitamin and
mineral supplementation: effect on the running performance of
trained athletes. Am. J. Clin. Nutr. 47:192–195, 1988.
28. WEIGHT, L. M., T. D. NOAKES,D.LABADARIOS,J.GRAVES,P.
JACOBS, and P. A. BERMAN. Vitamin and mineral status of trained
athletes including the effects of supplementation. Am. J. Clin.
Nutr. 47:186–191, 1988.
29. WELLER, E., P. BACHERT,H.MEINCK,B.FRIEDMANN,P.BARTSCH,
and H. MAIRBAURL. Lack of effect of oral Mg-supplementation on
Mg in serum, blood cells, and calf muscle. Med. Sci. Sports Exerc.
30:1584–1591, 1998.
498
Official Journal of the American College of Sports Medicine http://www.acsm-msse.org
