Magnesium and Potassium Balance

Increased serum magnesium concentrations lead to increased urinary loss of magnesium, which also may be exacerbated by stress hormone effects on the kidney. Urine magnesium losses in our study population were greater than normal (10 to 100 mg of magnesium/day [0.4 to 4.0 mmol]) even in the control group (270 mg of magnesium/day [20 mmol]), which had low-normal magnesium concentrations. The treatment group with high-normal to mildly increased serum magnesium concentrations lost nearly twice the magnesium in the urine (500 mg of magnesium/day) than the control group, suggesting that reabsorption of magnesium in the ascending loop of Henle is altered in critically ill patients. Although diuretics impair reabsorption of magnesium, urine magnesium concentrations in both groups varied dramatically and sporadically over time and were not linked chronologically to administration of diuretics, catecholamines, or other magnesium wasting therapy in either group. One patient lost 531 mg [21.2 mmol] of magnesium (6.2 mg/kg) in a single 6-hr period following Furosemide (40 mg iv). Another patient lost 435 mg [17.4] magnesium/6 hrs (5.4 mg/kg) without diuretics.

The osmotic diuresis seen with hyperglycemia is associated with increased urine loss of magnesium. Although the blood glucose concentrations were significantly different between the groups, the mean values for both groups were still at or below the normal threshold for glucosuria (160 to 190 mg/dL [9.0 to 11.4 mmol/L]). If there is an inherent defect in tubular transport that might alter this threshold, glucosuria could have occurred in members of the study population. Even if this circumstance was the case, however, the higher glucose concentrations that were present in the treatment group would have predisposed them to lose more magnesium than the control group. Thus, if the increased mean glucose concentration in the treatment group had any effect on the results, it was to diminish the difference between the groups rather than to falsely augment it.

Aggressive magnesium replacement has been advocated in the face of acute myocardial ischemia [29] and serious ventricular or supraventricular tachydysrhythmias. Although Speiser and Ganzoni’s study [29] of patients with proven acute myocardial infarction was retrospective, arrhythmias requiring treatment decreased from 46.3% in the control group to 13.4% in the group receiving 60 mmol of magnesium chloride over 2 days (p < .05). Recommendations included the administration of 13 g of magnesium sulfate intravenously over <6 hrs [29] . Flink [31] suggested a 4-g magnesium sulfate intravenous loading dose followed by 16 g infused over the subsequent 24 hrs in patients with seizures in the face of severe hypomagnesemia. In our study, critically ill patients with a high likelihood for magnesium deficiency were identified by the presence of hypokalemia. The dosage of magnesium sulfate was chosen because it reflected treatment that was more aggressive than our usual practice but was within the bounds of recommended treatment for hypomagnesemia. Administration of magnesium sulfate according to the study protocol led to a significantly more positive magnesium balance in the treatment group. Treatment patients received substantially more magnesium sulfate than the control group. Although it has been suggested that patients receiving total parenteral nutrition should receive 100 mg (4 mmol) of elemental magnesium per day to avoid hypomagnesemia [7] , our study suggests that severely stressed, critically ill patients have a daily need for elemental magnesium that is much higher than normal.

This study also showed that a significant improvement in potassium homeostasis occurred when serum magnesium concentrations were maintained at high-normal values. In the treatment group, potassium balance actually became positive, whereas the control group, maintained at a low-normal magnesium concentration, had a persistently negative potassium balance and had higher urine potassium losses despite virtually identical serum values and a trend toward greater potassium replacement. Of note, both groups showed a significant increase in potassium concentration during the 6 hrs after the first test dose. This increase likely reflects the surgical ICU’s usual policy of treating potassium concentrations of <4.0 mmol/L, such that most patients with hypokalemia sufficient to qualify for entry into the study were recruited shortly after ICU admission. In addition, it might suggest increased vigilance on the part of the nursing staff due to awareness of the patient's participation in the study, a factor that may have also encouraged administration of magnesium replacement in the control group. If this circumstance existed, the effect would be to diminish the differences between the groups.

Similarly, the higher glucose concentration found in the treatment group would have decreased rather than augmented the differences between the groups by increasing the tendency toward urine losses of potassium associated with osmotic diuresis. In addition, the higher osmolality resulting from hyperglycemia promotes loss of intracellular fluid, a subsequent increase in effective intracellular potassium concentrations,

Figure 3. Serum potassium (K+ ) (mmol/L) values and mmol excreted in the urine over time. Potassium balance represents the cumulative difference between mmol of potassium administered and that amount excreted in the urine. Open circles, treatment group; solid squares, control group. *p < .05; **p < .01; #p < .005. Values are expressed as mean ? SEM.

movement of potassium out of the cell, and delivery of higher potassium concentrations to the kidneys, further encouraging renal losses of potassium. Thus, if the higher glucose concentration in the treatment group had any effect on the results, it would have been more likely to have diminished the differences between the groups rather than to increase it.

Although creatinine clearance was not evaluated in this study, serum creatinine concentrations were within the normal range and were stable throughout the study. In addition, no patient was oliguric. Despite these values, however, it is possible, and even likely, that some patients had mild-to-moderate renal insufficiency. The extent of renal impairment cannot be ascertained from the available data but, due to the similarities between the groups with regard to severity of illness, diagnoses, and adjuvant kaliuretic and/or magnesium wasting therapies, there appear to be no glaring differences with regard to the impact of renal function on the results obtained.

Cardiac events are common in the critically ill population and survival to discharge after cardiac arrest occurs in <15% of in hospital patients with cardiac arrests. Potassium and magnesium depletion have been linked to cardiac dysrhythmias, coronary vasospasm, and cardiac arrest. Administration of magnesium has been shown to decrease the risk of dysrhythmias and death after myocardial infarction by 25% to 74% . Intracellular magnesium blocks outward potassium currents and provides the inward rectifying behavior of the ATP-dependent potassium channel, the potassium inward rectifier, the muscarinic potassium channel, and the voltage-dependent sodium channel. These rectifying potassium channels are involved in determining the plateau duration of the cardiac action potential, the rapid repolarization, and determination of the resting membrane potential. Thus, low intracellular magnesium concentrations alter the cardiac action potential and may predispose to propagation of dysrhythmias. Hypoxia and ischemia have been shown to cause an increase in efflux of potassium via ATP-sensitive potassium channels in mammalian ventricular myocytes, beginning within 15 to 30 secs of onset of ischemia. This flux of potassium causes a shortening of the action potential duration, and slowing of conduction, which may contribute to the development of reentrant ventricular tachydysrhythmias. Depression of the sodium potassium-ATPase pump may also contribute modestly through decreased influx of potassium under hypoxic and/or ischemic conditions. Activation of the N-methyl-D-aspartate receptor also results in calcium and sodium influx and utilizes intracellular magnesium and the N-methyl-D-aspartate receptor gated sodium and calcium channels to block and modulate the response. Although the numbers in this study were small, a trend toward decreased ectopy in the treatment group was noted compared with no change or an increase in the control group. A larger study population and a continuous, recorded electrocardiogram would be necessary to fully evaluate the effect of magnesium replacement on ectopy.

In conclusion, generous replacement of magnesium according to the study protocol was safe and effective. Both groups of critically ill patients studied had urine losses of magnesium that were much higher than normal.

TABLE 2 — Blood chemistries comparing treatment and control groups (mean ? SEM)

Treatment Control
Pre Post Pre Post
Glucose (mg/dL) 188 ? 20 a 185 ? 18 a 142 ? 8 139 ? 11
Arterial pH 7.40 ? 0.01 7.39 ? 0.02 7.39 ? 0.01 7.40 ? 0.01
Bicarbonate (mmol/L) 27 ? 1 28 ? 1 25 ? 1 27 ? 1
Calcium (mg/dL) 8.6 ? 0.2 8.4 ? 0.3 8.2 ? 0.2 8.4 ? 0.2
Phosphate (mg/dL) 3.1 ? 0.2 3.1 ? 0.3 3.9 ? 0.6 4.0 ? 0.5
BUN (mg/dL) 41 ? 6 41 ? 6 38 ? 6 39 ? 7
Creatinine (mg/dL) 1.0 ? 0.1 .9 ? 0.1 1.2 ? 0.2 1.2 ? 0.2
Pre, laboratory values measured upon entry into the study; Post, laboratory values measured upon completion of the study; BUN, blood urea nitrogen.
SI conversions: glucose mg/dL ? 0.056 = mmol/L; calcium mg/dL ? 0.25 = mmol/L; phosphate mg/dL ? 0.32 = mmol/L; BUN mg/dL ? 0.36 = mmol/L; creatinine mg/dL ? 0.88 = mumol/L.

a Significantly different from control (p < .05).

Figure 4. Total urine magnesium (Mg++ ) excretion (mg/48 hrs; 100 mg Mg = 4 mmol) as a function of the number of doses of diuretic administered/48 hrs. r = .206; r2 = .04; p = .48.

The study showed that a magnesium replacement strategy, which aimed to achieve a high-normal serum magnesium concentration, led to significantly greater retention of magnesium than the control group despite increased urine losses. In addition, administration of magnesium sulfate according to the treatment group protocol resulted in a positive potassium balance, whereas the control group, maintained at a low-normal magnesium concentration, exhibited a persistently negative potassium balance. The results of this study suggest that we need to reassess what we consider to be adequate serum magnesium concentrations in the critically ill. Furthermore, generous administration of magnesium sulfate (a low-risk approach) might obviate the high-risk approach of using potassium replacement infusions.


From the Departments of Anesthesiology (Dr. Hamill-Ruth) and Pharmacy (Dr. McGory), University of Virginia Health Sciences Center, Charlottesville, VA.Supported by Biomedical Research Support grant 5507-RR05431-29 from the National Institutes of Health.

Address requests for reprints to: Robin J. Hamill-Ruth, MD, Box 238, University of Virginia Health Sciences Center, Charlottesville, VA 22908.

Copyright 1996 by Williams & Wilkins

Critical Care Medicine

Volume 24 Number 1 January 1996

Copyright 1996 Williams & Wilkins

Robin J. Hamill-Ruth MD

Robb McGory PharmD

Special thanks are due to Carl Lynch, MD, PhD, for editorial assistance and the nursing staff of the surgical ICUs for their contributions to this study.


1. Chernow B, Bamberger S, Stroiko M, et al: Hypomagnesemia in patients in the postoperative intensive care unit. Chest 1989; 95:391-397

2. Ryzen E, Wagers PW, Singer FR, et al: Magnesium deficiency in a medical ICU population. Crit Care Med

3. Salem M, Munoz R, Chernow B: Hypomagnesemia in critical illness: A common and clinically important problem. Crit Care Clin 1991; 7:225-252

4. Rubeiz GJ, Thill-Baharozian M, Hardie D, et al: Association of hypomagnesemia and mortality in acutely ill medical patients. Crit Care Med 1993; 21:203-209

5. Fiaccadori E, Del Canale S, Coffrini E, et al: Muscle and serum magnesium in pulmonary intensive care unit patients. Crit Care Med 1988; 16:751-760

6. Hamill R, Robinson L, Moote C, et al: The efficacy and safety of potassium infusion therapy in hypokalemic critically ill patients. Crit Care Med 1991; 19:694-699

7. Whang R, Whang D, Ryan MP: Refractory potassium repletion: A consequence of magnesium deficiency. Arch Intern Med 1992; 152:40-44

8. Seelig M: Cardiovascular consequences of magnesium deficiency and loss: Pathogenesis, prevalence and manifestations–Magnesium and chloride loss in refractory potassium repletion. Am J Cardiol 1989; 63:4G-21G

9. Boyd J, Bruns D, DiMarco J, et al: Relationship of potassium and magnesium concentrations in serum to cardiac arrhythmias. Clin Chem 1984; 30:754-757

10. Boyd J, Bruns D, Wills M: Frequency of hypomagnesemia in hypokalemic states. Clin Chem 1983; 29:178-179

11. Munoz R, Khilnani P, Ziegler J, et al: Ultrafiltratable hypomagnesemia in neonates admitted to the neonatal intensive care unit. Crit Care Med 1994; 22:815-820

12. Fatemi S, Ryzen E, Flores J, et al: Effect of experimental human magnesium depletion on parathyroid hormone secretion and 1,25-dihydroxyvitamin D metabolism. J Clin Endocrinol Metab 1991; 73:1067-1072

13. Arsenian MA: Magnesium and cardiovascular disease. Prog Cardiovasc Dis 1993; 35:271-310

14. Altura B, Shang A, Altura B: Magnesium, hypertensive vascular disease, atherogenesis, subcellular compartmentation of Ca2+ and Mg2+ and vascular contractility. Miner Electrolyte Metab 1993; 19:323-336

15. Riggs J, Klingberg W, Flink E, et al: Cardioskeletal mitchondrial myopathy associated with chronic magnesium deficiency. Neurology 1992; 42:120-130

16. Classen HG, Scherb H, Scherb I: Invitro studien zur broncholysischen wirksamkeit von magnesium. Magnes Bull 1989; 11:80-84

17. Landon RA, Young EA: Role of magnesium in regulation of lung function. J Am Diet Assoc 1993; 93:674-677

18. Gunther T, Rebentisch E, Ising H, e

Print This Post Print This Post

Pages: 1 2