Anaesthesia and Intensive Care, Royal Adelaide Hospital/University
of Adelaide, North Terrace, Adelaide, Australia
The temporal relationship between the systemic and myocardial
concentrations of magnesium and some of its acute cardiovascular effects were examined after short i.v. infusion administration of
magnesium (30 mmol over 2 min) in five awake chronically instrumented sheep. Magnesium decreased mean arterial blood pressure and systemic vascular resistance (SVR) by 23 and 41% from baseline, respectively. These hemodynamic changes were consistent with magnesium producing primary reductions in SVR with partial heart rate (HR)-mediated compensation of blood pressure. Cardiac output and HR increased by 38 and 38% from baseline, respectively. Magnesium had little effect on
myocardial contractility, but substantially increased myocardial blood
flow (MBF, 77% above baseline) primarily due to direct myocardial
vasodilation. The peak arterial and coronary sinus serum magnesium
concentrations were 6.94 ± 0.26 (mean ± S.E.M.) and
6.51 ± 0.20 mM, respectively, at 2 min. Both arterial and
coronary sinus magnesium concentrations at the end of the study were
still more than 3 mM, whereas all the cardiovascular effects were back
to baseline. The myocardial kinetics of magnesium was consistent with
rapid equilibration of magnesium (half-life 0.4 min) with a small
distribution volume (71 ml) consistent with the extracellular space of
the heart. In conclusion, magnesium was shown to have a rapid
equilibration between the plasma/serum concentrations of magnesium and
its extracellular concentration in the myocardium. However, the primary
cardiovascular effect of magnesium (reductions in SVR) preceded its
extracellular concentrations, and was a direct function of its arterial
concentration. A "threshold" model for changes in SVR was preferred
when linked to the arterial magnesium concentration.
 |
Introduction |
Intravenous
magnesium has been increasingly used to treat and prevent many acute
cardiovascular diseases (Altura and Altura, 1985
; Nattel et al., 1991;
Hampton et al., 1994; Miller et al., 1995; Fawcett et al., 1999).
Functionally, magnesium can be regarded as a cardiovascular drug with
calcium antagonistic and antiadrenergic properties (James, 1992). Its
cardiovascular effects include direct and indirect dilatation of blood
vessels, an antiarrhythmic effect, and possibly myocardial depression.
Although it is known that there is a direct relationship between
magnesium plasma/serum concentration and cardiovascular effects at near
steady state (Friedman et al., 1987; James et al., 1987; Nakaigawa et
al., 1997), the relationship between the time courses of concentration and effect is poorly understood. Increased knowledge of this temporal relationship may provide a more rational basis for the design of acute
intravenous dose regimens of magnesium for the management of
cardiovascular symptoms.
There are several issues that need to be resolved before this can be
done. First, there is some ambiguity in the literature on the effects
of magnesium on some cardiovascular variables such as myocardial
contractility and myocardial blood flow, which may be due to
differences in magnesium dose and experimental conditions between
studies. Second, it is not known to what extent the plasma/serum concentrations of magnesium reflect the concentration of magnesium in
important organ systems mediating the cardiovascular effects (e.g.,
blood vessel walls for vasodilatation, and the myocardium for direct
effects on contractility and the ECG). Third, it is unclear whether the
cardiovascular effects of magnesium are mediated by its extracellular
or intracellular concentration, or a combination of both (Murphy et
al., 1991) in these organs.
We have previously examined the relationship between the myocardial
pharmacokinetics and dynamic of a number of drugs in sheep (Huang et
al., 1993a; Upton et al., 1996, 1999), and in this study applied these
methods to intravenous magnesium to address some of these issues. The
aims of the study were as follows. 1) To define the time course of some
of the cardiovascular effects of magnesium after an intravenous
infusion of 30 mmol of magnesium sulfate over 2 min to conscious,
instrumented sheep. Myocardial contractility and myocardial blood flow
were measured in a closed chested, conscious preparation without
concurrent drugs following a high dose of magnesium to deduce the
effect of magnesium alone on these cardiovascular variables. 2) To
define the time course of the systemic and myocardial concentrations of
magnesium, with the latter inferred from the concentration of magnesium
in coronary sinus blood leaving the heart. Modeling of myocardial
kinetics was used to test the hypothesis that the extracellular
magnesium concentration in the heart could be deduced from this
coronary sinus concentration (i.e., consistent with venous
equilibration of the extracellular space). 3) To examine the temporal
relationship between systemic and myocardial concentrations of
magnesium and key cardiovascular effects using kinetic-dynamic modeling.
| |
Materials and Methods |
General Experimental Preparation
The study was approved by the Animal Ethics Committee of the
University of Adelaide. The sheep were prepared in two steps. About 2 weeks before experimentation, sheep (2-3 years of age and
approximately 50 kg) were anesthetized with i.v. thiopental (1.5 g) and
the trachea intubated with a cuffed tracheal tube (9 mm i.d.; Sheridan
Catheter Corp., Argyle, NY). Anesthesia was maintained with 2%
halothane and 100% oxygen and end-expiratory carbon dioxide was
monitored using a capnograph (Cardiocap; Instrumentarium Corp,
Helsinki, Finland) and kept between 35 and 40 mm Hg.
The right femoral artery and vein were exposed via a groin incision.
Using the Seldinger technique (Runciman et al., 1984), a 7-French and a
9-French gauge catheter (Multi-purpose A1 catheter; Cook Australia,
Brisbane, Australia) were placed in the abdominal aorta. Through the
femoral vein, an 8.5-French introducer catheter (Biosensors
International Pty Ltd, Singapore) was placed in the inferior vena cava
(IVC). Through the introducer catheter, a 7.5-French multilumen
thermodilution catheter (Swan-Ganz, Biosensors International Pte Ltd)
was placed in the pulmonary artery. The position of the Swan-Ganz
catheter was confirmed by monitoring the pressure wave pattern with a
pressure transducer (model 4-327-I; Bell and Howell Inc., Pasadena, CA)
during catheter advancement (Runciman and Ludbrook, 1993).
Two days later, the sheep were anesthetized as described above for
probe placement and further catheterization using a modification of the
method reported previously (Huang et al., 1992). A left thoracotomy at
the 4th intercostal space and a pericardiotomy were performed to expose
the left main coronary artery and the pulmonary artery. Doppler flow
probes (Titronics Medical Instruments, Iowa City, IA) were placed
around the left main stem coronary artery (for measurement of an index
of left coronary blood flow) and pulmonary artery (for measurement of
cardiac output, CO). The probes were secured around the arteries with
cotton tape, which acted as a "cuff" around the arteries and
ensured a constant vessel caliber. The apex of the heart was stitched
with a 2-0 silk suture, a micro transducer (Codman MicroSensor; Johnson
& Johnson Professional, Inc., Raynham, Miami, FL) was placed 3 cm into the left ventricle through a 5-gauge needle via the apex of
heart and fixed securely with the suture. The hemiazygous vein, which
drains into the coronary sinus of sheep, was ligated outside the
pericardium, to ensure the coronary sinus contained pure effluent blood
from the myocardium. The leads of the probes and the micro transducer
were exteriorized although the chest incision and a subdermal tunnel.
The right jugular vein was exposed via a neck incision. A 7-French
gauge (B1; Cordis Corporation, Miami, FL) was place into the
coronary sinus to sample efferent blood from the myocardium. The
position of the catheters was confirmed under direct vision using a
fluoroscope with the injection of radio-opaque contrast (Conray 420;
May and Baker Ltd, Dagenham, UK) into the coronary sinus. All the
surgical procedures were carried out with using sterile technique and
all incisions were closed. The sheep were recovered from anesthesia and
housed in metabolic crates with free access to food and water. All
catheters were flushed at least every 2 days with heparinized (50 IU/ml) 0.9% saline.
Cardiovascular Measurements
During the experiments, the Doppler frequency shifts from both
the coronary artery and pulmonary artery Doppler flow probes were
recorded at a sampling rate of 1 Hz using a four-channel pulsed Doppler
flowmeter (Bioengineering, University of Iowa, 56 M.R.F, Iowa City, IA)
and an analog-to-digital card (MetraByte DAS 16-G2; MetraByte Corp.,
Taunton, Miami, FL) in a personal computer (Microbits 486-based
IBM compatible). These were used as indices of myocardial blood flow
and cardiac output, respectively. It has been shown that the left
coronary artery blood flow velocity was representative of flow in a
region corresponding to 77% of the heart (Huang et al., 1993b). The
pulmonary artery Doppler flow probes were calibrated in vivo by
correlating the Doppler shifts with cardiac output measured
intermittently using a thermodilution technique (Huang et al., 1992).
Mean arterial pressure (MAP) and central venous pressure (CVP) were
measured using a pressure transducer on an arterial catheter and an IVC
catheter, respectively. Left ventricular pressure, MAP, and CVP were
recorded using the same data acquisition system. The peak value of the
increasing rate of pressure rise of the left ventricle (LV
dP/dtmax) was calculated and used as an index of
myocardial contractility. Heart rate (HR) was also calculated from the
pressure wave of the left ventricle.
Myocardial oxygen consumption (MVO2) was
calculated as the product of myocardial blood flow (MBF, assuming a
baseline flow of 122 ml/min; Huang et al., 1992) and the
arteriocoronary sinus oxygen content gradient. Systemic vascular
resistance (SVR) was calculated as the difference between MAP and CVP
(mm Hg) over CO (l/min) and is reported in resistance units (mm Hg min
l
1).
Experimental Design
Studies were conducted in five sheep prepared as described
above. On the experimental day, the sheep remained in their metabolic crates with their weight supported by a sling to minimize sheep movement. A study was not commenced until at least 40 min after the
placement of the hemodynamic measurement devices, and the hemodynamic
measurements had been stable for approximately 10 min. After 5 min of
baseline hemodynamic measurements, 30 mmol of magnesium sulfate was
infused intravenously via the IVC catheter over 2 min (0.6 mmol/kg).
The start of the infusion was designated time zero. Arterial and
femoral venous blood samples (5 ml) were taken at 0, 1, 2, 4, 10, and
25 min for serum magnesium assay, and were assayed using a Ektachem 700 analyzer (Kodak GmbH, Stuttgart, Germany). Blood was also sampled from
arterial and coronary catheters at 0, 2, and 25 min after commencement
of administration of magnesium for blood gas analysis (ABL Radiometer
Medical A/S, Copenhagen, Denmark) to determine blood oxygen tension
(pO2), pH, blood carbon dioxide tension
(pCO2), and blood oxygen saturation
(SO2).
Data Analysis
Modeling of Myocardial Kinetics.
Hybrid modeling (Upton,
1996; Upton et al., 2000) was used to examine the ability of various
kinetic models to describe the observed magnesium concentrations in
coronary sinus blood. Models were constructed as a series of
differential equations with the Scientist for Windows software package
(version 2; Micromath Scientific Software, Salt Lake City, UT), and
were fitted to mean data for the five sheep after initial analysis
showed little interindividual variation in magnesium concentrations.
The measured arterial magnesium concentrations
(Cart) and blood flow entering the
heart were fitted to forcing functions (exponential and polynomial
functions, respectively) and these were used as the input functions for
the myocardial kinetic models. The measured coronary concentrations
(Ccs) were used to estimate the
parameters of the models by curve fitting. Curve fitting was by a
least-squares method based on the maximization of model selection
criteria (MSC) of the Scientist program (Upton, 1996; Upton et al.,
2000). Three models were examined as outlined below, where
Qh is myocardial blood flow, and
Vh is the apparent volume of magnesium
in the myocardium.
A single flow-limited compartment model:
|
(1)
|
A single flow-limited compartment with first order loss
model:
|
(2)
|
where kloss governs the rate of
the loss and has the units of volume per unit time.
A membrane-limited compartment model.
In this model,
"PS" is used to represent membrane permeability, and
Cdeep is the magnesium concentration
in the deep compartment of the myocardium with a volume given by
Vdeep:
|
(3)
|
A graphical representation of the three kinetic models is shown
in Table 2.
Kinetic-Dynamic Modeling.
SVR was considered the
primary cardiovascular parameter affected by magnesium. Four dynamic
models were examined that related the arterial or coronary sinus
magnesium concentrations (C) to this effect:
Linear Model.
A simple linear relationship between
concentration and effect, defined by the slope and intercept of a
line:
|
(4)
|
Linear Model with Delay.
As for the first model, but SVR was
related to magnesium concentration in a hypothetical effect compartment
whose time course of concentrations was delayed relative to the
measured concentration as given by the rate constant
keo. The rateconstant was constrained during fitting to be between 0 and 100 min1.
Ceff is the effect compartment
concentration:
|
(5)
|
A Linear Model with a Threshold.
This model was the same as
the linear model, but SVR was unchanged from baseline
(SVRbase) below a threshold (T)
concentration:
|
(6)
|
A Tolerance Model.
This model was adapted from that used by
Ekblom et al. (1993) to describe tolerance to morphine, and was such
that the effects of magnesium on SVR could become less with time, even
if the magnesium concentrations were held constant. This type of model
can empirically account for a variety of mechanisms of tolerance
(Mandema and Wada, 1995). Conceptually, the changes in SVR from
baseline (SVRbase) can be thought of as the net
result of a reduction attributed to the effect of magnesium
(SVRe), and an increase due to the development of
tolerance (SVRt):
|
(7)
|
The direct effect of magnesium was linearly related to
concentration:
The tolerance was related to concentrations in a hypothetical
tolerance compartment, which could be delayed relative to the measured
concentration, as given by the rate constant
kt0:
Statistical Analysis
To compare the effect of magnesium on MAP, SVR, CO, MBF, LV
dP/dtmax, HR, pH, pCO2,
pO2, SO2, and
MVO2, measurements were expressed as a percentage
of baseline, and then the mean and 95% confidence limits were
calculated. If the mean and its 95% confidence limits of a measured
value lay outside the 95% confidence limits of the baseline data, the
value was recorded as being statistically significant. Paired
t tests were used to compare the arterial and coronary sinus
magnesium concentrations. p < 0.05 was recorded as
statistically significant.
| |
Results |
Cardiovascular and Other Effects of Magnesium.
Magnesium
caused minor respiratory depression, as indicated by transient
reductions in arterial pO2 and
SO2, and increases in arterial
pCO2 (Table 1). The
increase of arterial pCO2 at 2 min was 29% above
baseline.
View this table:
[in this window]
[in a new window]
|
TABLE 1
Arterial and coronary sinus pH, pCO2, pO2, and
SO2 at baseline, 2 and 25 min in five sheep
Data are shown as mean (95% confidence limits) and letters indicate
significant difference when observed value and its expected 95%
confidence limits were outside the confidence limits of the baseline
data.
|
|
There were no significant changes in the time course of LV
dP/dtmax (Fig. 1A),
indicating that magnesium was not a myocardial depressant in this
experimental paradigm.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 1.
Time courses of dP/dtmax (A) and MBF (B).
Magnesium was administrated at time = 0 for 2 min. Data are shown
as the mean (symbols) and 95% confidence limits (dashed line). The
95% confidence limits of the baseline data are shown as the solid
lines. At any particular time, a statistically significant difference
occurs if the observed value and its 95% confidence intervals lay
outside of 95% confidence interval of the baseline period.
|
|
During and shortly after the infusion of magnesium, there was a
substantial increase in MBF from baseline, with a peak increase of 77%
at 1 min (Fig. 1B). However, there were no significant changes in
MVO2. The values of MVO2 at
baseline, and 2 and 25 min after the infusion (mean and 95% confidence
limits) were 4.4 (3-5.7), 3.9 (3.3-4.6), and 3.6 (2.7-4.6)
ml/min/100 g, respectively. Increased MBF in the face of unchanged
MVO2 is consistent with the observation that the
coronary sinus SO2 was significantly increased
above baseline at 2 min (Table 1). Overall, the increase in MBF caused
by magnesium appears to be due to a direct vasodilatory effect on the
coronary blood vessels, rather than secondary to cardiovascular changes
that increase myocardial work and therefore MVO2.
There were significant reductions in MAP and SVR from baseline (Fig.
2, A and B), with peak reductions of 23 and 41% for each, respectively, at 2 min. This was accompanied by an
increase in CO (Fig. 3A) and HR (Fig.
3B), with a peak increase of 38% at 2 min for both parameters. Stroke
volume (SV) was essentially unchanged (Fig. 3C). This pattern of
hemodynamic changes is consistent with magnesium-dependent reductions
in SVR, with partial compensation of MAP decreases via reflex increases
in heart rate. Increases in preload due to a transient fluid shift from
the arterial to venous circuit may also have contributed to the
increase in CO, although preload was not measured in the present study.
In subsequent kinetic dynamic analysis, a reduction in SVR was
considered to be the primary cardiovascular effect of magnesium, and it
was this effect that was related to the measured magnesium
concentrations.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 2.
Time courses of MAP (A) and SVR (B). Magnesium was
administrated at time = 0 for 2 min. Data are shown as the mean
(symbols) and 95% confidence limits (dashed line). The 95% confidence
limits of the baseline data are shown as the solid lines. At any
particular time, a statistically significant difference occurs if the
observed value and its 95% confidence intervals lay outside of 95%
confidence interval of the baseline period.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 3.
Time courses of cardiac output (CO) (A), HR (B), and
SV (C). Magnesium was administrated at time = 0 for 2 min. Data
are shown as the mean (symbols) and 95% confidence limits (dashed
line). The 95% confidence limits of the baseline data are shown as the
solid lines. At any particular time, a statistically significant
difference occurs if the observed value and its 95% confidence
intervals lay outside of 95% confidence interval of the baseline
period.
|
|
Myocardial Pharmacokinetics of Magnesium.
The observed
arterial and coronary sinus blood concentrations of magnesium are shown
in Fig. 4. There was a significant
difference between arterial and coronary sinus concentration at 1 min
after the start of the infusion. There was a general trend for uptake of magnesium into the heart (arterial > coronary sinus
concentrations) for approximately 10 min after the start of the
infusion.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 4.
Time course of observed mean arterial ( ) and
coronary sinus ( ) concentrations. Magnesium was administrated at
time = 0 for 2 min. Data are shown as mean and S.E.M. The mean
observed (symbols) and predicted coronary sinus concentrations for the
flow-limited model (solid line) are also in the inset.
|
|
The goodness of fit of the models of myocardial magnesium kinetics and
their parameter values are shown in Table
2. The flow-limited model was the best
fit of the data; the line of best fit to the coronary sinus magnesium
concentrations is shown in the inset of Fig. 4. The membrane-limited
model collapsed to a flow-limited model, as indicated by very small
values of PS and Vdeep, whereas the
flow-limited model with a first order loss returned a low and uncertain
value for the kloss.
View this table:
[in this window]
[in a new window]
|
TABLE 2
Goodness of fits of the model of magnesium kinetics in the myocardium
For MSC, the higher the number, the better the fit. The model
parameters were apparent distribution volume (Vh),
first order loss (kloss), membrane permeability
(PS), and deep compartment (Vdeep) in myocardium.
The data are shown as the mean and (S.D.).
|
|
The good fit of the flow-limited model is consistent with the concept
of a distribution volume of magnesium in the heart (approximately 71 ml) that is in equilibrium with coronary sinus blood. This volume can
be compared with the true volume of the region of the heart drained by
the coronary sinus catheter (233 g in sheep; Huang et al., 1993b), and
therefore represents approximately 30% of the total mass of the region
of the heart under study. This distribution volume can also be compared
with the baseline myocardial blood flow (122 ml/min; Huang et al.,
1992); the half-time for equilibration of the volume with serum was
therefore 0.40 min.
Pharmacokinetic-Pharmacodynamic Relationships.
The
parameters' values of the various dynamic models and their MSC values
are shown in Table 3. The linear dynamic
model produced an acceptable fit of the SVR data when linked to the arterial concentrations (Fig. 5), but was
less successful when linked to the coronary sinus concentrations (Table
3). Adding an effect delay to either did not improve the fit, and
produced estimates of keo that were
equal to the upper constraint (100 min1). This
value of keo produced time course of
measured concentration and effect compartment concentration that were
essentially identical. Overall, the data suggest that the arterial
magnesium concentrations were directly related to the reductions in SVR
with no time delay, and that the time course of the coronary sinus
concentrations lag behind both the time course of the arterial
concentrations and reductions in SVR.
View this table:
[in this window]
[in a new window]
|
TABLE 3
Goodness of fits of the dynamic models
For MSC, the higher the number, the better the fit. The class leader
for each site has been highlighted in bold.
|
|
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 5.
Kinetic-dynamic relationship for the linear effect
model. The solid circles are the observed time course of the arterial
magnesium concentration. Data are shown as mean and S.E.M. The open
circles are the observed time course of SVR. Data are shown as the mean
(symbols) and 95% confidence limits (dashed line). The solid line is
the line of best fit of the linear effect model linked to the arterial
magnesium concentrations.
|
|
Close inspection of Fig. 5 will show that the greatest
discrepancy between the line of best fit for the linear model and the SVR data occurs for the last time point (25 min), although this is also
the value of SVR with the widest 95% confidence intervals. The
application of threshold and tolerance models to the data were an
attempt to account for this discrepancy, which is consistent with SVR
returning to near baseline while the magnesium concentrations remained
elevated (Figs. 2B and 4). The threshold model was preferred when
linked to the arterial concentration, whereas the tolerance model was
preferred when linked to the coronary sinus concentrations, but
returned an uncertain value for the tolerance rate constant (ktol). This suggests that the
threshold dynamic model was suitable for describing the relationship
between the arterial magnesium concentration and SVR.
| |
Discussion |
Cardiovascular Effects of Magnesium.
The pattern of
hemodynamic changes observed was consistent with peripheral
vasodilatation, hypotension, and a reflex increase in CO, predominantly
due to an increase in HR. The use of an awake preparation avoided any
anesthetic depression of autonomic reflexes or myocardial function that
may complicate the interpretation of some earlier studies (Friedman et
al., 1987). The changes observed were similar to those previously
reported in studies using lightly anesthetized baboons (James et al.,
1987), although decreases in HR were more common in studies where
magnesium concentrations reached supratherapeutic concentrations
(Nakaigawa et al., 1997).
Myocardial Contractility.
There was no evidence of myocardial
depression, as determined by LV dP/dtmax, despite
magnesium concentrations reaching around 7 mM, although it is possible
a small direct myocardial depressant effect may have been countered by
a reflex sympathetic response to hypotension. This is consistent
with much of the published literature, which generally shows minimal
cardiac depression at therapeutic concentrations (in the order of 2-3
times baseline levels). Critelli et al. (1977) reported that bolus
administration of magnesium in humans produced only small decreases in
indices of myocardial contractility in patients with mild
myocardiosclerosis, but more significant decreases in patients with
impaired cardiac function (New York Heart Association class III-IV). In
lightly anesthetized baboons, magnesium produced minimal myocardial
depression, and even at very high magnesium concentrations a decrease
in CO seemed to be HR mediated (James et al., 1987). Nakaigawa et al. (1997) reported that magnesium produced dose-dependent decreases in LV
dP/dtmax in anesthetized dogs. This depression
was, however, minor at blood concentrations similar to those achieved
in the current study, but became marked with blood concentrations of around 9 to 14 mg/dl. In vitro, Bass et al. (1958) demonstrated using a
modified Langendorff heart preparation that low concentrations of
magnesium (0.025-2.5 mg) did not affect myocardial contractility, whereas higher doses (25-250 mg) temporarily depressed contractility. In contrast, Friedman et al. (1987) found magnesium reduced LV dP/dtmax by around 30% with magnesium
concentrations similar to those in the current study; chloralose
anesthesia may have been a contributing factor.
The use of awake animals in the present study meant that some
respiratory depression and mild hypercarbia was induced. Although it is
possible that hypercarbia influenced the data in the current study, it
is known that even very large increases in pCO2
produce minimal or no changes in CO, LV dP/dtmax,
and HR (Larrieu et al., 1978; Foex and Ryder, 1979; van den Bos et al.,
1979), suggesting any influence in the current study was minimal.
Myocardial Blood Flow.
Although an increase in MBF would be
expected in the present study, considering the large decrease in SVR,
MBF is also influenced by the work performed by the heart. This in turn
is influenced (broadly) by SV, MAP, and HR. Following the
administration of magnesium in the present study, MAP decreased whereas
HR increased. The overall work of the heart therefore appeared
unchanged, in agreement with the unchanged MVO2.
The increase in MBF in the face of unchanged MVO2
is consistent with the increase in coronary sinus oxygen content.
Overall, the data suggest that magnesium produced coronary hyperemia
due to direct vasodilatation of coronary blood vessels. This is
consistent with the work of Scott et al. (1961) who reported that MBF
velocity was almost doubled, and coronary vascular resistance was
decreased, after intracoronary infusion of magnesium in the dog heart,
and in vitro studies showing, magnesium to have direct coronary
vasodilating effects (Bass et al., 1958; Altura and Altura, 1984,
1985). In contrast, Nakaigawa et al. (1997) reported no change in MBF
or in CO, but MVO2 and oxygen extraction both decreased.
In the current study, the small decrease in hemoglobin saturation due
to magnesium-induced respiratory depression was unlikely to have
affected MBF, but it is possible that the concurrent mild hypercarbia
may have had a small influence. Hypercarbia has been shown to produce
increases in MBF (Foex and Ryder, 1979; van den Bos et al., 1979), and
these increases have been shown to exceed increases in
MVO2 (Foex and Ryder, 1979; van den Bos et al.,
1979). However these increases have been associated with marked
hypercarbia (over 70 mm Hg), and it seems unlikely that the
pCO2 of around 45 mm Hg in the current study
contributed more than a small degree to the large (77%) increase in
MBF velocity. This is supported by the observations from pilot studies
using the present experimental preparation where it was found that
increasing inspired CO2 from 0 to 10% had a
minimal effect on MBF.
Myocardial Pharmacokinetics of Magnesium.
Magnesium ions
equilibrate slowly across cell membranes (probably via ion channels).
Intracellular concentrations are a function of complex buffering of
magnesium ions with ATP and other molecules within the cell (Raftos et
al., 1999). The electrochemical gradient across the cell membrane
drives the flux of extracellular magnesium into the cell, but the
intracellular concentrations of free magnesium are much less than
expected by this process. The low intracellular concentration at
equilibrium is achieved by relatively slow active transport of
magnesium out the cell via a
Na+-Mg2+ antiports
(Flatman, 1991).
The kinetic modeling supports the notion that this active transport of
magnesium out of the cell is sufficient to constrain the distribution
volume of exogenous magnesium to the extracellular space. The
distribution volume was found to be 30% of the mass of the region of
the heart studied, which includes the blood. The extracellular volume
of heart without blood in rat has been reported to be 15 to 18% of the
total volume, depending on the markers used (Makos et al., 1998). In
addition to the contribution of blood to the distribution volume, there
may be some differences in the binding of magnesium between the serum
and interstitial fluid that contributes to the larger volume observed
in the present study. Certainly, the data suggest that there is very
rapid equilibration of magnesium between serum and the extracellular
space, and that the magnesium concentration in venous serum emerging
from the heart is representative of its concentration in the
extracellular space. In most circumstances where magnesium is
administered more slowly than in the present study, it would be
expected that the total serum concentrations of magnesium are
representative of the total magnesium concentration in the
extracellular space.
Pharmacokinetic-Pharmacodynamic Relationships.
Magnesium is
known to modify the properties of a number of ion channels in the cell
membrane. Although some channels are affected by the intracellular
concentration of magnesium, others (including its calcium channel
blocking effects) are affected by its extracellular concentration
(Murphy et al., 1991; Bara et al., 1993). It would seem likely that the
reductions in SVR caused by magnesium are a function of its calcium
channel blocking effects relaxing vascular smooth muscle. However, the
data suggest that the "global" extracellular concentrations of
magnesium, as represented by the effluent concentrations from the
heart, do not determine this primary cardiovascular effect of
magnesium. The fact that this effect was better related to the arterial
concentrations of magnesium suggest that magnesium is acting directly
on "proximal" segments of the microvasculature, presumably the
arterioles responsible for regulating vascular resistance. Capillary
permeability theory dictates that the proximal segments of the exchange
microvasculature equilibrate first with the afferent arterial blood,
whereas distal segments equilibrate with efferent venous blood (Stec
and Atkinson, 1981).
An interesting finding in the study was that both the arterial and
coronary sinus magnesium concentrations at the end of the study were
more than 3 mM, at which time all the cardiovascular variables had
returned to baseline. A threshold effect was tested, as Bolan et al.
(1985) demonstrated a lack of hemodynamic effects with blood magnesium
concentrations below 4 mM in pregnant sheep. Alternatively, an acute
tolerance to the effects of magnesium would explain the same
observation. The mechanism of tolerance could take a number of forms,
most of which are mathematically equivalent to the tolerance model
investigated here. For example, serum magnesium exists in two forms:
bound to covalent ligands (i.e., plasma proteins) and free ionized
magnesium. The free magnesium fraction in the whole blood is about 59 to 71% (Brookes and Fry, 1993; Ising et al., 1995; Hafen et al.,
1996). Total serum magnesium was measured in the present study,
whereas only free ionized magnesium is thought to have biological
activity. Tolerance could be accounted for by progressive, slow
conversion of free to bound magnesium during the course of the study.
In our own laboratory, we have found that the vasodilatory effect of
magnesium infused directly into a muscle bed diminishes with time
(Zheng et al., 2000).
Implications from the study are first that there is rapid equilibration
between the plasma/serum concentrations of magnesium and its
extracellular concentration in the myocardium. It can be concluded that
monitoring arterial concentrations of magnesium is therefore
representative of its extracellular concentrations in the heart.
Second, the primary cardiovascular effects of magnesium (reductions in
SVR) were initially related to these arterial concentrations. However,
important cardiovascular effects may not be directly related to these
concentrations after longer time periods (
25 min).
We thank Dr. K. Rowland and staff of the Institute of Medical
and Veterinary Science, South Australia for performing the magnesium assay.
This research was supported by grants from National Health and Medical
Research Council of Australia.
IVC, inferior vena cava;
CO, cardiac output;
MAP, mean arterial blood pressure;
CVP, central venous pressure;
LV dP/dtmax, maximum positive rate of change of left
ventricular pressure;
HR, heart rate;
MVO2, myocardial
oxygen consumption;
MBF or Qh, myocardial
blood flow;
SVR, systemic vascular resistance;
pO2, blood
oxygen tension;
pCO2, blood carbon dioxide tension;
SO2, blood oxygen saturation;
MSC, model selection
criteria;
SV, stroke volume;
Cart, arterial
magnesium concentration;
Ccs, coronary sinus
magnesium concentration;
Vh, apparent
distribution volume of magnesium in the myocardium.
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
Abstract
Introduction
