AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 7, NUMBER 1 January 1990
MAGNESIUM SULFATE PHARMACOKINETICS PREGNANT CAPRA HIRCUS MODEL Gary D. V Hankins, M.D., Russell R. Snyder, M.D., John C. Hauth, M.D., Larry C. Gilstrap, HI, M.D., Frank H. Wians, Jr., Ph.D., and Adrian E Van Dellen, D.V.M.
ABSTRACT
Magnesium sulfate (MgSO4-7H2O) remains the mainstay of antiseizure therapy for preeclampsia in the United States and, although few question its efficacy in this capacity, unanimity of opinion regarding the exact mechanism and site of action of magnesium sulfate is lacking. !~5 It is well established that magnesium sulfate reduces the sensitivity of the motor end-plate to applied acetylcholine and thereby decreases the amplitude of the motor end-plate potential. It is also postulated, however, that magnesium ions act centrally on the cerebral cortex to prevent seizures and that the level of cortical suppression is related to the plasma magnesium concentration.1'6 To date, no such correlation between plasma and cortical or cerebrospinal fluid (CSF) magnesium concentrations has been consistently found. The ideal route of magnesium sulfate administration, intravenous versus intramuscular is also controversial.1"5 With vasospasm the hallmark of preeclampsia or eclampsia, it is puzzling that equivalent plasma magnesium levels have been reported using either 5 gm of intramuscularly administered magnesium sulfate every 4 hours or 8 gm administered intravenously (2 gm/hr) over the same time interval.5
Downloaded by: National University of Singapore. Copyrighted material.
A comparison of standard clinical regimens for the administration of magnesium sulfate for treatment of preeclampsia was performed in the pregnant goat model. The regimen of intravenous or intramuscular load and intramuscular maintenance championed by Pritchard was found to yield higher maternal serum levels through the first 4 hours of treatment compared with intravenous load with intravenous maintenance therapy (p < 0.05); however, neither regimen affected the concentration of magnesium ion in the cerebrospinal fluid. Urinary excretion of magnesium and passage into the amniotic fluid were also evaluated for each route of administration and neither accounted for the disparity in serum concentrations noted during the first 4 hours of magnesium therapy.
These results are, indeed, opposite of predictions based on the pathophysiology of the disease process. Moreover, when the drug is solely administered intravenously to eclamptic women, seizures recur at a rate 2.7 times greater than seen with Pritchard's combination intravenous or intramuscular protocol.7"9 Higher central nervous system magnesium concentrations with maintenance intramuscular administration of magnesium sulfate might explain these apparent discrepancies, especially inasmuch as through the first 4 hours of treatment total magnesium administered would equal 14 gm with the intramuscular regimen versus 12 gm with a 4 gm intravenous load and 2 gm/hr maintenance protocol. Accordingly, the purpose of this study was to determine the concentration of magnesium ion in maternal serum, urine, and CSF and in fetal amniotic fluid (AF) following either intravenous or intramuscular routes of administration. MATERIALS AND METHODS
Fifteen healthy pregnant goats (Capra hircus), with a mean body weight of 31.3 kg (range, 19.5 to
Department of Obstetrics and Gynecology, Wilford Hall, United States Air Force Medical Center, Lackland Air Force Base, Texas The opinions expressed in this manuscript are those of the authors and not necessarily those of the United States Air Force or the Department of Defense. Reprint requests: Dr. Hankins, Department of Obstetrics and Gynecology, Wilford Hall, USAF Medical Center/SGHPG, Lackland Air Force Base, TX 78236-5300 Copyright © 1990 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.
49
40) and an average body surface area of 1.07 m2, were used in this study. All animals were in the last one third of their gestation. With the exception of water, the goats were fasted for 12 hours prior to the studies. This experimental protocol was approved by the Investigational Review Board and Animal Use and Care Committee of Wilford Hall USAF Medical Center and by the Surgeon General's Office. Anesthesia
Anesthesia was initiated by intramuscular ketamine (15 to 20 mg/kg), and endotracheal intubation was performed to ensure airway patency and to prevent aspiration. A surgical level of anesthesia was maintained using 0.5 to 1.0 volumes percent halothane, 50% nitrous oxide, and 50% oxygen until the animals were instrumented. Thereafter, light anesthesia was maintained by titration of the halothane to the lowest effective concentration (0.3 to 1.5 vol. %). Instrumentation
After establishment of anesthesia, the femoral artery and vein were cannulated with a 20 gauge, 2 inch Silastic catheter and an 8 F venous introducer (Cordis Corp., Miami, Florida). A 7 F flow-directed pulmonary artery catheter was then positioned in the pulmonary artery via waveform pattern recognition. All animals maintained stable peripheral and central pressures, cardiac output, and arterial and mixed venous blood gases during the experiment. Additionally, fetal heart rates were periodically monitored visually by sonography and remained stable during the experiment. A 20 gauge Silastic catheter with multiple distal apertures was inserted into the maternal spinal canal percutaneously so that CSF could be sampled continuously. Catheter dead space was predetermined to ensure that samples represented the true dynamics of the CSF. Urine output was quantified hourly by an indwelling Foley catheter. Study Design
Two treatment regimens were studied with six animals in each group. Magnesium sulfate was administered on a weight-adjusted basis to correspond to currently used clinical regimens. Specifically, the weight adjustment assumed the average pregnant woman to weigh 70 kg. An animal weighing 35 kg would receive a 50% dose of the magnesium regimen tested, and one weighing 28 kg would receive a 40% dose of the regimen tested. The regimens tested corresponded to a 10 gm intramuscular and 4 gm intravenous load with a 5 gm intramuscular maintenance dose every 4 hours (group I), and a 4 gm intravenous load followed by a maintenance dose of 2 gm/hr given as a continuous intravenous infu50 sion (group II). A third group consisted of three
January 1990
pregnant animals in whom all procedures were done identically to the test animals but who received no magnesium sulfate. Intravenous fluid administration was controlled by infusion pumps and verified by timed volutrol measurements. All infused fluids consisted of 5 % dextrose in water at a rate of 1.79 ml/ kg/hr, the equivalent of 125 ml/hr in the pregnant woman. Those animals receiving continuous infusions of magnesium sulfate received the drug in the specified weight-adjusted volume of fluid and did not receive additional fluid. The total duration of the experiment was 8 hours with the first hour required for instrumentation and the second hour for equilibration and acquisition of baseline maternal serum, urine, CSF, and AF specimens. Maternal plasma and AF specimens were analyzed for sodium, chloride, magnesium, total and ionized calcium, and creatinine concentrations. Maternal CSF and urine were analyzed for magnesium and calcium ion concentrations. Urine electrolytes, creatinine, and magnesium concentrations were also determined. All biochemical parameters were determined by our accredited reference clinical laboratory. All equipment was calibrated and quality controlled as per manufacturer and standard specifications. Magnesium concentrations (milligrams percent units) were measured with the Dupont Automatic Clinical Analyzer. Means, standard deviations, and analysis of variances and covariances of repeated measures were performed using a standard statistical computer software package (BMDP Simple Data Description and Data Management Package, Department of Biomathematics, University of California, Los Angeles). Significance was defined as p < 0.05.
RESULTS
Although animals were randomly assigned to the treatment groups, their weights were different (34.6 kg versus 28.1 kg, p < 0.05), which resulted in a significantly larger total dose of magnesium sulfate (p tol lotai 1 A Ionized 6 hours Total Ionized
1 L . \ — \ ./
4.5 ± 0.5
4.8 ± 0.4
12.2 ± 1.8 4.5 ± 0.4
11.8 ± 1.9 4.8 ± 0.4
Downloaded by: National University of Singapore. Copyrighted material.
Table 1. Magnesium Metabolism According to Route of Administration*
groups across time, or between treated and control groups. Total urinary excretion of magnesium is listed in Table 1. Although group I excreted more magnesium than group II, whether measured as the ion concentration or as a percentage of the total amount administered, the difference was not statistically significant. The concentration of urinary calcium increased across time in both treated and control animals, but failed to achieve statistical significance. Urinary calcium excretion was not affected by the route of the magnesium administration. The urinary creatinine concentration of treated animals did not differ from control animals nor did it change across time. DISCUSSION
Magnesium sulfate is standard therapy for both the prevention and treatment of eclamptic seizures;1-4'8 however, both its mechanism of action and ideal route of administration are subject to intense debate.8"11 Intramuscular magnesium sulfate for the treatment of preeclampsia or eclampsia, as described and validated by Pritchard, 78 has been unparalleled relative to a favorable outcome for both the mother and her fetus. Pritchard encountered only five failures in the treatment of 245 eclamptic women. Accordingly, we tested Pritchard's 10 gm intramuscular and 4 gm intravenous load with 5 gm intramuscular maintenance every 4 hours as de-
Comparison of Total Dose of Magnesium Sulfate in Each Group with the Serum Concentration and Ratio of Serum Concentration to Total Dose of Magnesium Salt 2 Hours
Total dose magnesium sulfate (gm) Serum level (mg/dl) Ratio mg/dl/gm
6.92 7.9 1.14
4 Hours
6 Hours
II
1
II
3.2 4.9 1.53
6.92 6.8 0.98
4.81 4.8 1.00
9.39 7.4 0.79
6.42 5.5 0.86
51
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 7, NUMBER 1
January 1990
Table 5. Maternal CSF Magnesium (Mean ± SD) and Calcium Ion Concentrations (Mean ± SD) and Percent Change Relative to Basal Levels for Both Treated and Control Animals
Basal (mg/dl) Treatment 2 hours Percent change 4 hours Percent change 6 hours Percent change
Group II
Control
Magnesium
Calcium
Magnesium
Calcium
Magnesium
Calcium
-t-
0.4
8.1 ± 0.3
2.4 ± 0.4
9.7 ± 2.7
2.3 ± 0.2
10 ± 0.9
2.8 ± 2.9 4 2.8 ± 1.2 6 2.9 ± 0.2 7
8.5 ± 0.9
2.6 ± 0.4
8.1 ± 4.2
2.7 ± 0.3
10 ± 0.9
2.7
3
9
-16
8.5 ± 2.9
2.7 ± 0.4
9.8 ± 2.9
3
14
8.6 ± 3.0
2.8 ± 0.6
4
17
scribed for women with severe disease (group I). The other regimen tested (group II) was the 4 gm intravenous load and 2 gm/hr continuous intravenous infusion as reported by Sibai et al.2>3 This particular intravenous regimen was chosen because it has been shown to achieve maternal serum levels comparable to those achieved with the intramuscularly administered drug. Furthermore, doses greater than 2 gm/ hr have not achieved widespread use due to fear of toxicity,12 and in actual clinical practice many physicians use a 1 gm/hr dosage schedule even though as few as 2% of patients have been reported to achieve therapeutic magnesium levels.2 In the current investigation we found the intramuscular regimen to yield higher maternal serum magnesium ion concentrations at both 2 and 4 hours into treatment (p < 0.001 and < 0.005, respectively) relative to the intravenous regimen. Moreover, for the entire 6 hours of treatment evaluated by us, the intramuscular group sustained a greater increase relative to baseline than the intravenous group (94 versus 81%). Although these results are contrary to what one might predict based on the pathophysiology of preeclampsia with widespread vasospasm, they are essentially identical to those reported by Sibai et al.3 They found that during the first 3 hours of therapy the intramuscular regimen produced serum magnesium levels that were consistently higher than those achieved with the 2 gm/hr intravenous regimen. Perhaps this simply reflects the difference in total magnesium administered through the first 4 hours, 14 versus 12 gm in the two regimens tested by us based on use in clinical practice. Indeed, as shown in Table 3, the actual plasma concentration of magnesium, when expressed relative to grams of salt given, was greater at 2 hours with the intravenous regimen. This, then, suggests that the lower initial levels, at 1 to 4 hours of therapy, are simply reflective of inadequate initial intravenous therapy. To define further the distribution of magnesium sulfate as a function of route of administration and to explain the differences in serum concentrations, we also sequentially monitored maternal urinary excretion of magnesium ion as well as the accumulation of magnesium ion in the CSF and AF. The 52 sequential concentration of magnesium in urine, the
2
9.7 ± 2.8 0
15
2.5 ± 0.2 6
2.5 ± 0.2 8.6
0
9.3 ± 1 -6.7 9.8 ± 0.8 -1.7
percent excreted at each time period, and the total percent excreted were not significantly different relative to the route of administration and did not account for the differences observed in serum levels. Furthermore, no significant change in the maternal CSF magnesium level occurred at any time interval for either treatment group relative to untreated control animals. Lastly, analysis of AF revealed no increase in magnesium ion concentration over the course of this study, regardless of the treatment regimen. From prior work, however, it is well established that the fetus is a "sink" for magnesium.5 The appearance in AF, however, requires maternal absorption, placental passage, fetal renal excretion, and, finally, fetal urination, and the duration of this study may have been insufficient for some of these processes to occur. Nonetheless, through 6 hours we found no increase in amniotic fluid magnesium sufficient to account for the difference in maternal serum concentrations, especially at 2 hours into therapy when the largest difference in percent change was observed. Thus, neither urinary excretion, passage into the AF, nor passage into the CSF adequately explains the differences in maternal serum magnesium levels achieved during the early hours of treatment between the intramuscular and intravenous regimens. In the present study there were no significant changes in total calcium either across time during magnesium administration or between treatment groups. Thus, the antagonistic effects of free or ionized calcium appear an unlikely explanation for any differences in efficacy relative to administration routes. Failure to demonstrate an increase in the CSF magnesium concentration or a change in the CSF calcium concentration, regardless of the route of administration, again raises the question of the mechanism and site of action of magnesium sulfate. Although one can question the validity of extrapolation of animal data to humans, others have reported CSF magnesium levels in both preeclamptic and eclamptic women.8 Levels with magnesium therapy were only 2.65 mEq/liter compared with 2.35 ± 0.15 mEq/L in untreated pregnant control subjects.8 Additionally, in one woman treated continuously with magnesium sulfate for 7 days and with documented
Downloaded by: National University of Singapore. Copyrighted material.
Group
MAGNESIUM SULFATE PHARMACOKINETICS/Hankins, Snyder, Hauth, et al.
SUMMARY
That magnesium sulfate will prevent or stop an eclamptic seizure is irrefutable.1617 However, the exact mechanism by which it does so is still unknown. Any centrally induced effects do not appear to be brought about by acute increases in the CSF magnesium concentration. Intramuscular administration of magnesium sulfate supplemented with an initial 4 gm intravenous load yields higher maternal
serum levels during the initial 2 to 4 hours of treatment compared with that attained with only continuous intravenous administration. The most likely explanation for the higher serum levels is the higher dose, especially throughout the initial 2 to 4 hours of treatment.
REFERENCES
1. Pritchard JA, McDonald PC: Williams Obstetrics, 17th ed. New York: Appleton-Century-Crofts, 1985, p 547 2. Sibai BM, Lipshitz J, Andersen GD, et al: Reassessment of intravenous MgSO4 therapy in preeclampsia-eclampsia. Obstet Gynecol 57:199-202, 1981 3. Sibai BM, Graham JM, McCubbin JH: A comparison of intravenous and intramuscular magnesium sulfate regimens in preeclampsia. Am J Obstet Gynecol 150: 728733, 1984 4. Sibai BM, McCubbin JH, Andersen GD, et al: Eclampsia I. Observations from 67 recent cases. Obstet Gynecol 58: 609-613, 1981 5. Pritchard JA: The use of the magnesium ion in the management of eclamptogenic toxemias. Surg Gynecol Obstet 100:131-140, 1955 6. Borges LF, Gucer G: Effect of magnesium on epileptic foci. Epilepsia 19:81-91, 1978 7. Pritchard JA, Cunningham FG, Pritchard SA: The Parkland Memorial Hospital protocol for treatment of eclampsia: Evaluation of 245 cases. Am J Obstet Gynecol 148:951 — 963, 1984 8. Pritchard JA: The use of magnesium sulfate in preeclampsia-eclampsia. J Reprod Med 23:107-113, 1979 9. Sibai BM, Spennato JA, Watson DL, et al: Effect of magnesium sulfate on electroencephalographic findings in preeclampsia-eclampsia. Obstet Gynecol 64:261-266, 1984 10. Donaldson JO: Neurology of Pregnancy. Philadelphia: WB Saunders, 1978, pp 229-232 11. Ferris TF: Toxemia and hypertension, In Medical Complications During Pregnancy, 2nd ed. Burrow GN, Ferris TF (eds): Philadelphia: WB Saunders, 1982, p 82 12. McCubbin J, Sibai BM, Abdella TN, et al: Cardiopulmonary arrest due to acute maternal hypermagnesaemia. Lancet 1:1058, 1981 13. Bohdanecky Z, Weiss T: The electrophysiological manifestation of the narcotic state evoked by magnesium sulfate. Act Nerv Super 6:42, 1964 14. Dordoni DF: Ricerche elettroencefalografiche sulla cosiddetta narcosi da magnesio e sull' antagonismo calciomagnesio. Arch Sci Biol 42:59, 1958 15. Koontz WL, Reid KH: Effect of parenteral magnesium sulfate on penicillin induced seizure foci in anesthetized cats. Am J Obstet Gynecol 153:96-99, 1985 16. Storchheim F: Status epilepticus treated by magnesium sulphate injected intravenously. JAMA 101:1313—1314, 1933 17. Lazard EM: A preliminary report on the intravenous use of magnesium sulphate in puerperal eclampsia. Am J Obstet Gynecol 9:178-188, 1925
Downloaded by: National University of Singapore. Copyrighted material.
sustained serum levels of 7.6 to 13.2 mg/dl, the CSF level was only 4.2 mg/dl. From this study, it was concluded that the CSF magnesium concentration remained almost totally independent of the maternal and fetal plasma level.8 It would appear, then, that the blood-brain barrier is relatively impervious to magnesium and, based on the measurements of CSF levels in humans, is maintained intact even in preeclamptic women. In addition to failure of magnesium sulfate to achieve levels in the CSF equivalent to those considered therapeutic in serum, Bohdanecky and Weiss,13 as well as Dordeni,14 have reported that magnesium infusion failed to change the spontaneous electroencephalographic pattern or arousal responses in animals. In addition, Sibai et al9 reported that magnesium sulfate did not alter abnormal electroencephalographic patterns in either preeclamptic or eclamptic women, even at proven "therapeutic" serum levels. The latter human data are in marked contradistinction to the results Borges and Gucer6 obtained in a number of animal species. They showed a reduction in electroencephalographic neuronal burst firing and interictal spike generation in animals made epileptic by topical application of penicillin G to the cerebral cortex. The degree of spike suppression increased in parallel with the plasma magnesium concentration and decreased as the magnesium level decreased, Koontz and Reid,15 however, found no significant effects of magnesium even at a peak plasma concentration of 11.7 ± 2.0 mg/dl on penicillin-induced epileptiform spikes in anesthetized cats relative to control animals. Importantly, with the penicillin model used by Borges and Gucer, as well as by Koontz and Reid, both the number and frequency of epileptiform spikes rapidly decays across time even in the absence of specific therapy, suggesting that Borges and Gucer's findings may have been coincidental and not causal.
MAGNESIUM SULFATE PHARMACOKINETICS PREGNANT CAPRA HIRCUS MODEL Gary D. V Hankins, M.D., Russell R. Snyder, M.D., John C. Hauth, M.D., Larry C. Gilstrap, HI, M.D., Frank H. Wians, Jr., Ph.D., and Adrian E Van Dellen, D.V.M.
ABSTRACT
Magnesium sulfate (MgSO4-7H2O) remains the mainstay of antiseizure therapy for preeclampsia in the United States and, although few question its efficacy in this capacity, unanimity of opinion regarding the exact mechanism and site of action of magnesium sulfate is lacking. !~5 It is well established that magnesium sulfate reduces the sensitivity of the motor end-plate to applied acetylcholine and thereby decreases the amplitude of the motor end-plate potential. It is also postulated, however, that magnesium ions act centrally on the cerebral cortex to prevent seizures and that the level of cortical suppression is related to the plasma magnesium concentration.1'6 To date, no such correlation between plasma and cortical or cerebrospinal fluid (CSF) magnesium concentrations has been consistently found. The ideal route of magnesium sulfate administration, intravenous versus intramuscular is also controversial.1"5 With vasospasm the hallmark of preeclampsia or eclampsia, it is puzzling that equivalent plasma magnesium levels have been reported using either 5 gm of intramuscularly administered magnesium sulfate every 4 hours or 8 gm administered intravenously (2 gm/hr) over the same time interval.5
Downloaded by: National University of Singapore. Copyrighted material.
A comparison of standard clinical regimens for the administration of magnesium sulfate for treatment of preeclampsia was performed in the pregnant goat model. The regimen of intravenous or intramuscular load and intramuscular maintenance championed by Pritchard was found to yield higher maternal serum levels through the first 4 hours of treatment compared with intravenous load with intravenous maintenance therapy (p < 0.05); however, neither regimen affected the concentration of magnesium ion in the cerebrospinal fluid. Urinary excretion of magnesium and passage into the amniotic fluid were also evaluated for each route of administration and neither accounted for the disparity in serum concentrations noted during the first 4 hours of magnesium therapy.
These results are, indeed, opposite of predictions based on the pathophysiology of the disease process. Moreover, when the drug is solely administered intravenously to eclamptic women, seizures recur at a rate 2.7 times greater than seen with Pritchard's combination intravenous or intramuscular protocol.7"9 Higher central nervous system magnesium concentrations with maintenance intramuscular administration of magnesium sulfate might explain these apparent discrepancies, especially inasmuch as through the first 4 hours of treatment total magnesium administered would equal 14 gm with the intramuscular regimen versus 12 gm with a 4 gm intravenous load and 2 gm/hr maintenance protocol. Accordingly, the purpose of this study was to determine the concentration of magnesium ion in maternal serum, urine, and CSF and in fetal amniotic fluid (AF) following either intravenous or intramuscular routes of administration. MATERIALS AND METHODS
Fifteen healthy pregnant goats (Capra hircus), with a mean body weight of 31.3 kg (range, 19.5 to
Department of Obstetrics and Gynecology, Wilford Hall, United States Air Force Medical Center, Lackland Air Force Base, Texas The opinions expressed in this manuscript are those of the authors and not necessarily those of the United States Air Force or the Department of Defense. Reprint requests: Dr. Hankins, Department of Obstetrics and Gynecology, Wilford Hall, USAF Medical Center/SGHPG, Lackland Air Force Base, TX 78236-5300 Copyright © 1990 by Thieme Medical Publishers, Inc., 381 Park Avenue South, New York, NY 10016. All rights reserved.
49
40) and an average body surface area of 1.07 m2, were used in this study. All animals were in the last one third of their gestation. With the exception of water, the goats were fasted for 12 hours prior to the studies. This experimental protocol was approved by the Investigational Review Board and Animal Use and Care Committee of Wilford Hall USAF Medical Center and by the Surgeon General's Office. Anesthesia
Anesthesia was initiated by intramuscular ketamine (15 to 20 mg/kg), and endotracheal intubation was performed to ensure airway patency and to prevent aspiration. A surgical level of anesthesia was maintained using 0.5 to 1.0 volumes percent halothane, 50% nitrous oxide, and 50% oxygen until the animals were instrumented. Thereafter, light anesthesia was maintained by titration of the halothane to the lowest effective concentration (0.3 to 1.5 vol. %). Instrumentation
After establishment of anesthesia, the femoral artery and vein were cannulated with a 20 gauge, 2 inch Silastic catheter and an 8 F venous introducer (Cordis Corp., Miami, Florida). A 7 F flow-directed pulmonary artery catheter was then positioned in the pulmonary artery via waveform pattern recognition. All animals maintained stable peripheral and central pressures, cardiac output, and arterial and mixed venous blood gases during the experiment. Additionally, fetal heart rates were periodically monitored visually by sonography and remained stable during the experiment. A 20 gauge Silastic catheter with multiple distal apertures was inserted into the maternal spinal canal percutaneously so that CSF could be sampled continuously. Catheter dead space was predetermined to ensure that samples represented the true dynamics of the CSF. Urine output was quantified hourly by an indwelling Foley catheter. Study Design
Two treatment regimens were studied with six animals in each group. Magnesium sulfate was administered on a weight-adjusted basis to correspond to currently used clinical regimens. Specifically, the weight adjustment assumed the average pregnant woman to weigh 70 kg. An animal weighing 35 kg would receive a 50% dose of the magnesium regimen tested, and one weighing 28 kg would receive a 40% dose of the regimen tested. The regimens tested corresponded to a 10 gm intramuscular and 4 gm intravenous load with a 5 gm intramuscular maintenance dose every 4 hours (group I), and a 4 gm intravenous load followed by a maintenance dose of 2 gm/hr given as a continuous intravenous infu50 sion (group II). A third group consisted of three
January 1990
pregnant animals in whom all procedures were done identically to the test animals but who received no magnesium sulfate. Intravenous fluid administration was controlled by infusion pumps and verified by timed volutrol measurements. All infused fluids consisted of 5 % dextrose in water at a rate of 1.79 ml/ kg/hr, the equivalent of 125 ml/hr in the pregnant woman. Those animals receiving continuous infusions of magnesium sulfate received the drug in the specified weight-adjusted volume of fluid and did not receive additional fluid. The total duration of the experiment was 8 hours with the first hour required for instrumentation and the second hour for equilibration and acquisition of baseline maternal serum, urine, CSF, and AF specimens. Maternal plasma and AF specimens were analyzed for sodium, chloride, magnesium, total and ionized calcium, and creatinine concentrations. Maternal CSF and urine were analyzed for magnesium and calcium ion concentrations. Urine electrolytes, creatinine, and magnesium concentrations were also determined. All biochemical parameters were determined by our accredited reference clinical laboratory. All equipment was calibrated and quality controlled as per manufacturer and standard specifications. Magnesium concentrations (milligrams percent units) were measured with the Dupont Automatic Clinical Analyzer. Means, standard deviations, and analysis of variances and covariances of repeated measures were performed using a standard statistical computer software package (BMDP Simple Data Description and Data Management Package, Department of Biomathematics, University of California, Los Angeles). Significance was defined as p < 0.05.
RESULTS
Although animals were randomly assigned to the treatment groups, their weights were different (34.6 kg versus 28.1 kg, p < 0.05), which resulted in a significantly larger total dose of magnesium sulfate (p tol lotai 1 A Ionized 6 hours Total Ionized
1 L . \ — \ ./
4.5 ± 0.5
4.8 ± 0.4
12.2 ± 1.8 4.5 ± 0.4
11.8 ± 1.9 4.8 ± 0.4
Downloaded by: National University of Singapore. Copyrighted material.
Table 1. Magnesium Metabolism According to Route of Administration*
groups across time, or between treated and control groups. Total urinary excretion of magnesium is listed in Table 1. Although group I excreted more magnesium than group II, whether measured as the ion concentration or as a percentage of the total amount administered, the difference was not statistically significant. The concentration of urinary calcium increased across time in both treated and control animals, but failed to achieve statistical significance. Urinary calcium excretion was not affected by the route of the magnesium administration. The urinary creatinine concentration of treated animals did not differ from control animals nor did it change across time. DISCUSSION
Magnesium sulfate is standard therapy for both the prevention and treatment of eclamptic seizures;1-4'8 however, both its mechanism of action and ideal route of administration are subject to intense debate.8"11 Intramuscular magnesium sulfate for the treatment of preeclampsia or eclampsia, as described and validated by Pritchard, 78 has been unparalleled relative to a favorable outcome for both the mother and her fetus. Pritchard encountered only five failures in the treatment of 245 eclamptic women. Accordingly, we tested Pritchard's 10 gm intramuscular and 4 gm intravenous load with 5 gm intramuscular maintenance every 4 hours as de-
Comparison of Total Dose of Magnesium Sulfate in Each Group with the Serum Concentration and Ratio of Serum Concentration to Total Dose of Magnesium Salt 2 Hours
Total dose magnesium sulfate (gm) Serum level (mg/dl) Ratio mg/dl/gm
6.92 7.9 1.14
4 Hours
6 Hours
II
1
II
3.2 4.9 1.53
6.92 6.8 0.98
4.81 4.8 1.00
9.39 7.4 0.79
6.42 5.5 0.86
51
AMERICAN JOURNAL OF PERINATOLOGY/VOLUME 7, NUMBER 1
January 1990
Table 5. Maternal CSF Magnesium (Mean ± SD) and Calcium Ion Concentrations (Mean ± SD) and Percent Change Relative to Basal Levels for Both Treated and Control Animals
Basal (mg/dl) Treatment 2 hours Percent change 4 hours Percent change 6 hours Percent change
Group II
Control
Magnesium
Calcium
Magnesium
Calcium
Magnesium
Calcium
-t-
0.4
8.1 ± 0.3
2.4 ± 0.4
9.7 ± 2.7
2.3 ± 0.2
10 ± 0.9
2.8 ± 2.9 4 2.8 ± 1.2 6 2.9 ± 0.2 7
8.5 ± 0.9
2.6 ± 0.4
8.1 ± 4.2
2.7 ± 0.3
10 ± 0.9
2.7
3
9
-16
8.5 ± 2.9
2.7 ± 0.4
9.8 ± 2.9
3
14
8.6 ± 3.0
2.8 ± 0.6
4
17
scribed for women with severe disease (group I). The other regimen tested (group II) was the 4 gm intravenous load and 2 gm/hr continuous intravenous infusion as reported by Sibai et al.2>3 This particular intravenous regimen was chosen because it has been shown to achieve maternal serum levels comparable to those achieved with the intramuscularly administered drug. Furthermore, doses greater than 2 gm/ hr have not achieved widespread use due to fear of toxicity,12 and in actual clinical practice many physicians use a 1 gm/hr dosage schedule even though as few as 2% of patients have been reported to achieve therapeutic magnesium levels.2 In the current investigation we found the intramuscular regimen to yield higher maternal serum magnesium ion concentrations at both 2 and 4 hours into treatment (p < 0.001 and < 0.005, respectively) relative to the intravenous regimen. Moreover, for the entire 6 hours of treatment evaluated by us, the intramuscular group sustained a greater increase relative to baseline than the intravenous group (94 versus 81%). Although these results are contrary to what one might predict based on the pathophysiology of preeclampsia with widespread vasospasm, they are essentially identical to those reported by Sibai et al.3 They found that during the first 3 hours of therapy the intramuscular regimen produced serum magnesium levels that were consistently higher than those achieved with the 2 gm/hr intravenous regimen. Perhaps this simply reflects the difference in total magnesium administered through the first 4 hours, 14 versus 12 gm in the two regimens tested by us based on use in clinical practice. Indeed, as shown in Table 3, the actual plasma concentration of magnesium, when expressed relative to grams of salt given, was greater at 2 hours with the intravenous regimen. This, then, suggests that the lower initial levels, at 1 to 4 hours of therapy, are simply reflective of inadequate initial intravenous therapy. To define further the distribution of magnesium sulfate as a function of route of administration and to explain the differences in serum concentrations, we also sequentially monitored maternal urinary excretion of magnesium ion as well as the accumulation of magnesium ion in the CSF and AF. The 52 sequential concentration of magnesium in urine, the
2
9.7 ± 2.8 0
15
2.5 ± 0.2 6
2.5 ± 0.2 8.6
0
9.3 ± 1 -6.7 9.8 ± 0.8 -1.7
percent excreted at each time period, and the total percent excreted were not significantly different relative to the route of administration and did not account for the differences observed in serum levels. Furthermore, no significant change in the maternal CSF magnesium level occurred at any time interval for either treatment group relative to untreated control animals. Lastly, analysis of AF revealed no increase in magnesium ion concentration over the course of this study, regardless of the treatment regimen. From prior work, however, it is well established that the fetus is a "sink" for magnesium.5 The appearance in AF, however, requires maternal absorption, placental passage, fetal renal excretion, and, finally, fetal urination, and the duration of this study may have been insufficient for some of these processes to occur. Nonetheless, through 6 hours we found no increase in amniotic fluid magnesium sufficient to account for the difference in maternal serum concentrations, especially at 2 hours into therapy when the largest difference in percent change was observed. Thus, neither urinary excretion, passage into the AF, nor passage into the CSF adequately explains the differences in maternal serum magnesium levels achieved during the early hours of treatment between the intramuscular and intravenous regimens. In the present study there were no significant changes in total calcium either across time during magnesium administration or between treatment groups. Thus, the antagonistic effects of free or ionized calcium appear an unlikely explanation for any differences in efficacy relative to administration routes. Failure to demonstrate an increase in the CSF magnesium concentration or a change in the CSF calcium concentration, regardless of the route of administration, again raises the question of the mechanism and site of action of magnesium sulfate. Although one can question the validity of extrapolation of animal data to humans, others have reported CSF magnesium levels in both preeclamptic and eclamptic women.8 Levels with magnesium therapy were only 2.65 mEq/liter compared with 2.35 ± 0.15 mEq/L in untreated pregnant control subjects.8 Additionally, in one woman treated continuously with magnesium sulfate for 7 days and with documented
Downloaded by: National University of Singapore. Copyrighted material.
Group
MAGNESIUM SULFATE PHARMACOKINETICS/Hankins, Snyder, Hauth, et al.
SUMMARY
That magnesium sulfate will prevent or stop an eclamptic seizure is irrefutable.1617 However, the exact mechanism by which it does so is still unknown. Any centrally induced effects do not appear to be brought about by acute increases in the CSF magnesium concentration. Intramuscular administration of magnesium sulfate supplemented with an initial 4 gm intravenous load yields higher maternal
serum levels during the initial 2 to 4 hours of treatment compared with that attained with only continuous intravenous administration. The most likely explanation for the higher serum levels is the higher dose, especially throughout the initial 2 to 4 hours of treatment.
REFERENCES
1. Pritchard JA, McDonald PC: Williams Obstetrics, 17th ed. New York: Appleton-Century-Crofts, 1985, p 547 2. Sibai BM, Lipshitz J, Andersen GD, et al: Reassessment of intravenous MgSO4 therapy in preeclampsia-eclampsia. Obstet Gynecol 57:199-202, 1981 3. Sibai BM, Graham JM, McCubbin JH: A comparison of intravenous and intramuscular magnesium sulfate regimens in preeclampsia. Am J Obstet Gynecol 150: 728733, 1984 4. Sibai BM, McCubbin JH, Andersen GD, et al: Eclampsia I. Observations from 67 recent cases. Obstet Gynecol 58: 609-613, 1981 5. Pritchard JA: The use of the magnesium ion in the management of eclamptogenic toxemias. Surg Gynecol Obstet 100:131-140, 1955 6. Borges LF, Gucer G: Effect of magnesium on epileptic foci. Epilepsia 19:81-91, 1978 7. Pritchard JA, Cunningham FG, Pritchard SA: The Parkland Memorial Hospital protocol for treatment of eclampsia: Evaluation of 245 cases. Am J Obstet Gynecol 148:951 — 963, 1984 8. Pritchard JA: The use of magnesium sulfate in preeclampsia-eclampsia. J Reprod Med 23:107-113, 1979 9. Sibai BM, Spennato JA, Watson DL, et al: Effect of magnesium sulfate on electroencephalographic findings in preeclampsia-eclampsia. Obstet Gynecol 64:261-266, 1984 10. Donaldson JO: Neurology of Pregnancy. Philadelphia: WB Saunders, 1978, pp 229-232 11. Ferris TF: Toxemia and hypertension, In Medical Complications During Pregnancy, 2nd ed. Burrow GN, Ferris TF (eds): Philadelphia: WB Saunders, 1982, p 82 12. McCubbin J, Sibai BM, Abdella TN, et al: Cardiopulmonary arrest due to acute maternal hypermagnesaemia. Lancet 1:1058, 1981 13. Bohdanecky Z, Weiss T: The electrophysiological manifestation of the narcotic state evoked by magnesium sulfate. Act Nerv Super 6:42, 1964 14. Dordoni DF: Ricerche elettroencefalografiche sulla cosiddetta narcosi da magnesio e sull' antagonismo calciomagnesio. Arch Sci Biol 42:59, 1958 15. Koontz WL, Reid KH: Effect of parenteral magnesium sulfate on penicillin induced seizure foci in anesthetized cats. Am J Obstet Gynecol 153:96-99, 1985 16. Storchheim F: Status epilepticus treated by magnesium sulphate injected intravenously. JAMA 101:1313—1314, 1933 17. Lazard EM: A preliminary report on the intravenous use of magnesium sulphate in puerperal eclampsia. Am J Obstet Gynecol 9:178-188, 1925
Downloaded by: National University of Singapore. Copyrighted material.
sustained serum levels of 7.6 to 13.2 mg/dl, the CSF level was only 4.2 mg/dl. From this study, it was concluded that the CSF magnesium concentration remained almost totally independent of the maternal and fetal plasma level.8 It would appear, then, that the blood-brain barrier is relatively impervious to magnesium and, based on the measurements of CSF levels in humans, is maintained intact even in preeclamptic women. In addition to failure of magnesium sulfate to achieve levels in the CSF equivalent to those considered therapeutic in serum, Bohdanecky and Weiss,13 as well as Dordeni,14 have reported that magnesium infusion failed to change the spontaneous electroencephalographic pattern or arousal responses in animals. In addition, Sibai et al9 reported that magnesium sulfate did not alter abnormal electroencephalographic patterns in either preeclamptic or eclamptic women, even at proven "therapeutic" serum levels. The latter human data are in marked contradistinction to the results Borges and Gucer6 obtained in a number of animal species. They showed a reduction in electroencephalographic neuronal burst firing and interictal spike generation in animals made epileptic by topical application of penicillin G to the cerebral cortex. The degree of spike suppression increased in parallel with the plasma magnesium concentration and decreased as the magnesium level decreased, Koontz and Reid,15 however, found no significant effects of magnesium even at a peak plasma concentration of 11.7 ± 2.0 mg/dl on penicillin-induced epileptiform spikes in anesthetized cats relative to control animals. Importantly, with the penicillin model used by Borges and Gucer, as well as by Koontz and Reid, both the number and frequency of epileptiform spikes rapidly decays across time even in the absence of specific therapy, suggesting that Borges and Gucer's findings may have been coincidental and not causal.
