Journal of Internal Medicine 1999; 246: 373-378
Abstract. Rob PM, Dick K, Bley N, Seyfert T, Brinckmann CH, Höllriegel V, Friedrich HJ, Dibbelt L, Seelig MS (Medizinische Universität zu Lübeck, Freie Universität Berlin, Germany and University of North Carolina, Chapel Hill, USA). Can one really measure magnesium deficiency using the short-term magnesium loading test? J Intern Med 1999 246: 373-378.
Objective. To compare a 1-h-version of a magnesium-loading-test (MLT) designed for outpatients in healthy controls with the 8-h standard; to establish the test in patients after renal transplantation prone to develop magnesium (Mg) deficiency; to correlate femur Mg-concentration and percentage retention of the given load.
Design. Comparison of mean values from healthy controls with respective from the literature; a prospective, randomized, controlled 4-month study; an intra-individual correlation of Mg-serum values and loading-test data with femur-Mg concentrations.
Setting. One centre study in a medical university; outpatients from the transplant unit; inpatients from the orthopedic unit.
Subjects. Twenty-four healthy controls aged 36.7 ± 7.4 years; 34 patients after renal transplantation (46.5 ± 14.3 years); 41 patients with hip replacement therapy (63.9 ± 18.6 years).
Intervention. Baseline Mg values were measured by atomic absorption spectroscopy (AAS) in serum and urine. An intravenous Mg load with 0.1 mmol Mgaspartate hydrochloride per kilogram bodyweight was given during 1 h. In 24 h-urine, the amount of excreted Mg was measured by AAS and the percentage retention of the given load calculated according to the formula: 1 - [Mg 24 h-urine/Mg test dose] X 100. Femur Mg was measured by AAS in a piece of the femur neck. Patients after renal transplantation were randomized after the first Mg load to either obtain daily 5 mmol Mg-aspartate hydrochloride per kilogram bodyweight, or placebo. Four months later a second loading-procedure was performed.
Main outcome measure. Serum Mg, percentage retention of the given Mg load (%Ret) and femur Mg concentration.
Results. Mean serum Mg values were within the normal range. In controls, %Ret was -18 ± 21 and not different from the literature. In the first MLT after renal transplantation, %Ret was 47 ± 43. In patients under Mg medication it decreased significantly to 16 ± 26, but was 58 ± 27 in the placebo group. Femur Mg concentration was 62.6 ± 20.9 mmol kg-1 dry substance and the corresponding %Ret was 14 ± 28 with r = - 0.7093.
Conclusion. The short-term version of the MLT is as good as the standard and was easily applied in outpatients. The indication from the good correlation between bone-Mg and %Ret and a marked decrease in %Ret in patients after Mg medication was that one can really measure magnesium deficiency.
Keywords: magnesium deficiency, magnesium loading test, magnesium retention.
Magnesium (Mg) is involved in numerous biochemical and physiological processes; its deficiency has been shown to play a role in a variety of clinically important diseases . It is crucial to be able to detect magnesium deficiency [2-4]. However, the most widely used test - serum magnesium - is not a reliable index of the magnesium status. Only a minority of patients with abnormal serum Mg levels is identified by physician initiated order of the serum Mg level . Hypomagnesemia, as reflected by low serum Mg, indicates Mg deficiency, but intracellular Mg depletion - involving many tissues - can exist despite normal serum Mg levels [6, 7]. There is poor correlation between serum Mg and tissue pools of Mg, except for the interstitial fluid . Less than 1% of total body Mg is in the blood, and only 0.3% is in the serum, yet most clinical data derive from serum sampling and analysis . Of the 1% of Mg that is in the blood, most is in the cells and platelets, and analysis of Mg in erythrocytes, lymphocytes, mono-nuclear blood cells and platelets have yielded important data, as have muscle and bone biopsies. The techniques, however, are too time consuming, too invasive and/or too expensive for routine measurement [9, 10] and the results have been conflicting [9, 11]. Furthermore, there is little evidence for a dynamic equilibrium amongst body tissues . The magnesium-loading-test (MLT) was first used as a means to verify Mg depletion induced in two healthy volunteers who had submitted to dietary Mg deficiency on a diet that provided only 1.1 mEq of Mg a day for three to four weeks . The two young men retained 25 and 45% of the intravenously administered Mg at the end of the study. The first clinical report showing the utility of the MLT was in patients who experienced gastrointestinal Mg loss . In that study, the Mg was administered either intramuscularly or intravenously, and retention of 20% or more was considered evidence of deficiency. There has been growing interest amongst cardiologists in detecting Mg deficiency by means of the MLT [15, 16]. As many as three quarters of 100 elderly patients with congestive heart failure, hypertension, and/or diabetes mellitus were found to be Mg deficient by the MLT . In a larger study of patients, predominantly with cardiovascular diseases, but with representation of diabetics, alcoholics and some with gastrointestinal disorders, who were given intravenous Mg loading over 8 hours , the mean percentage retention of Mg ranged between 22% and 54%. A reference range for Mg retention in 88 healthy men and women between 18 and 66 years of age, who were given 30 mmol of Mg i.v. during 8 h and urine collection from start of infusion for 24 h was given by Gullestad et al. . Their mean Mg retention was 6.3 ± - 10.3% and the 0.025 and 0.975 fractiles were -19.5% and 27.5% of the loading dose, significantly less than the same investigative group found in the patients population . Because even 8 hours infusion of the Mg load is unsuitably long for outpatients, we have undertaken studies of 1-hour Mg infusion in healthy volunteers  and in patients who had undergone renal transplantation with (Mg wasting) immunosuppressive drug therapy . Presented here are additional details on those studies, with the addition of comparative retention and bone Mg level data on hip replacement patients with gastrointestinal disorders, who were given intravenous Mg loading over 8 hours , the mean percentage retention of Mg ranged between 22% and 54%. A reference range for Mg retention in 88 healthy men and women between 18 and 66 years of age, who were given 30 mmol of Mg i.v. during 8 h and urine collection from start of infusion for 24 h was given by Gullestad et al. . Their mean Mg retention was 6.3 ± - 10.3% and the 0.025 and 0.975 fractiles were -19.5% and 27.5% of the loading dose, significantly less than the same investigative group found in the patients population . Because even 8 hours infusion of the Mg load is unsuitably long for outpatients, we have undertaken studies of 1-hour Mg infusion in healthy volunteers  and in patients who had undergone renal transplantation with (Mg wasting) immunosuppressive drug therapy . Presented here are additional details on those studies, with the addition of comparative retention and bone Mg level data on hip replacement patients.
Subjects and methods
Study of the feasibility of a 1-hour infusion test load of Mg, was undertaken to determine percentage retention - for assay of the Mg status of three groups of test subjects - according to the recommendations of the local ethics committee. All subjects gave written informed consent. Group 1 comprised 24 healthy individuals, who received no medication; Group 2 was made up of 34 patients 4 months after renal transplantation, who were being immunosuppressed with ciclosporine, and who were prospectively randomized after the first loading dose of Mg to receive: (i) 5 mg kg-1 bodyweight per day magnesium aspartate hydrochloride (Magnesiocard; Verla Pharm, Tutzing, Germany) by mouth; or (ii) placebo, for 4 months - after which the MLT was repeated. Group 3 was made up of 41 patients who received a MLT after hip replacement and who had a 1-cm bone slice cut from the middle of the removed femur neck, which after washing in cold NaCl 0.9% was stored at -20 °C for analysis for bone Mg.
The MLT was performed between 8:00 h and 10:00 h. Exclusion criteria were AV-block, heart rate 5560 beats min-1, blood pressure =110/ 70 mm mercury, pulmonary congestion or dyspnoe, serum creatinine concentration >200 µmol L-1 and proteinuria >500 mg day-1. After emptying the urinary bladder, 0.1 mmol Mg per kilogram bodyweight as the aspartate hydrochloride (Magnesiocard; Verla Pharm) in 500 mL isotonic saline were infused over 1 hour. Blood samples were drawn without stasis 60 min after starting the infusion, to measure the peak serum magnesium produced by the load. Blood pressure and electrocardiogramms were recorded twice during the infusion. Urine was collected in plastic bottles containing 15 mL 10% HC1, from the start of the infusion for 24 h. The urine as measured to the nearest of 50 mL, and an aliquot of the entire sample was sent by outpatients to our laboratory by mail. Subjects were informed of Mg-rich foods that were to be avoided during the collection period. Mg retention was calculated from the 24-h urinary Mg excretion and expressed as a percentage of the amount of the Mg test dose according to the formula :
I - [ Mg in 24 h urine/ Mg test dose] x 100
Dietary intake, fecal excretion and basal urinary magnesium output were ignored. The freeze dried bone was analysed for Mg by dissolving 0.2-0.5 g of the bone by heating with 10 mL 10 N HNO3. The solution was given into a 100-mL beaker which was filled up with aqua bidestillata. Then, 200 µL were taken and put into 4 mL 0.1% La/0.1 N HC1 which was measured by atomic absorption spectrophotometry  (AAS 1100; Perkin Elmer, Lindau, Germany). Urine Mg was also determined by AAS. Data are given as mean ± SD. Percent retention of the magnesium-load is given as the median, and although it is statistically not allowed to give the standard deviation in case of percentage data, we did so according to the literature, for better comparison. Statistical analysis was done by ANOVA, Scheffe's test and Pearson's correlation. *Indicates a statistical significant difference with a P = 0.05.
Neither in healthy individual, nor in patients, was there significant change in blood pressure, heart frequency, or electrocardiogram during the MLT infusion. Some individuals reported a sensation of warmth. Demographic data and values obtained before the Mg-load are given in Table 1. Serum-Mg was significantly lower in patients but their mean level was within the normal range in all groups. Basic Mg-excretion and serum albumin concentration did not differ in the groups. One hour after the infusion began, the maximum serum-Mg concentration was not different between healthy controls and patients (Table 2). After the load, 24 h urine volume did not differ significantly in the patients (1878 ± 769 mL) from the controls (1730 ± 610 mL). In healthy subjects, the percentage of Mg retained was -18 ± 21, which is comparable to that obtained in a larger study: - 19 ± 27.5 . In patients 4 months after renal transplantation, 47 ± 43% of the Mg load was retained, which is the mean plus three standard deviations of the control group. Four months later, the percentage retention decreased towards normal values in patients who were Mg supplemented (n = 16), but decreased in the placebo group (n = 16, P = 0.001). Complete data were available for analysis from 21 patients after hip replacement; they retained 14 ± 28% of the Mg load, and the Mg concentration in the stored slice of the femur neck was 62.65 ± 20.97 mmol Mg per kilogram dry substance. The correlation coefficient between both was r = - 0.7093 [P < 0.001; Fig. 1], whereas there was no significant correlation between serum levels and bone Mg content (r = 0.2217).
No adverse reaction was encountered using the short-term version of MLT: administration of 0.1 mmol Mg/kg as Mg aspartate hydrochloride in 500 mL saline over 1 hour. Outpatients had no problems related to the sampling and measuring of the urine volume and the mailing of an aliquot. In healthy individuals, the percentage retention of Mg was comparable with that obtained with 30 mmol (as MgSO4 in 500 mL saline) over 8 hours . The negative values of retention means that the amount of excreted magnesium is greater than the Mg load. This results from the fact that we did not take into account baseline Mg excretion which is possible , but needs a second 24 h urine sampling opposing the aim of this study to simplify the test. The broad range can be the result of the fact that dietary Mg intake was not standardized, in view of the difficulty in enforcing strict nutritional restrictions in outpatients, and the fact that there are thus resultant broad day-to-day variations in Mg excretion. Peak serum Mg concentrations were lower in this study than in that reported by Gullestadt et al. [4; 1.85 ± 0.54 mmol L-1 Mg]. This may be advantageous, since the higher the maximum magnesium concentration, the greater the possibility of overcoming maximum capacity for magnesium reabsorption in the renal tubular system (Tm Mg) to give false negative results because Tm Mg is close to the upper level of the normal serum magnesium concentration . Groups 2 and 3 were also normomagnesemic regarding the mean value, but at the lower limit of the Mg reference value in our laboratory (1.15-0.75 mmol L-1). Patients after renal transplantation were outpatients in whom Mg depletion might have developed due to renal Mg wasting  related to a defective renal tubular system because of ischemia and the nephrotoxicity of various drugs especially ciclosporine [7, 10, 16]. The loading-test was applied to answer the question of whether there is magnesium deficiency or not. Despite a normal serum Mg level, we found a high percentage retention of the Mg load which was greater than the mean plus 3 SD of the control population indicating magnesium deficiency . The validity of the long-term Mg loading test version of Gullestad et al.  to detect Mg deficiency was demonstrated by a decrease in magnesium retention in a second loading test in the same individuals some months later, after continuous oral magnesium medication . In our study, too, the percentage retention of the Mg-load turned towards normal values after 4 months of Mg application. However, in patients to whom no Mg was given there was even a higher retention rate of the given Mg load, giving evidence for the assumption of magnesium deficiency. Finally, we could show that the results obtained with the short term MLT reflects, very well, bone Mg content, which is the most relevant Mg store. Based on the above given data we conclude that one can really measure Mg deficiency in outpatients using this version of the MLT.
1 Wacker WEC, Parisi AF. Magnesium metabolism. N Engl J Med 1968; 278: 658-63;712-7;772-6.
2 Elin RJ. Assessment of magnesium status. Clin Chem 1987; 33: 311-4.
3 Alfrey AC, Miller NL, Butkus D. Evaluation of body magnesium stores. J Lab Clin Med 1974; 84: 153-62.
4 Gullestad L, Midtvedt K, Dolva LO, Norseth J. Kjeskshus J. The magnesium loading test: reference values in healthy subjects. Scand J Clin Lab Invest 1994; 54: 23-31.
5 Whang R, Ryder KW. Frequency of hypomagnesemia and hypermagnesemia. Requested vs routine. J Am Med Assoc 1990; 263: 3063-4.
6 Dyckner T, Wester PO. The relation between extra- and intracellular electrolytes in patients with hypokalemia and/ or diuretic treatment. Acta Med Scand 1978; 204: 269-82.
7 Rob PM, Bley N, Dick K, Rohwer J. Sack K. Magnesium deficiency after renal transplantation and Cyclosporine treatment despite normal serum magnesium concentration detected by a modified magnesium-loading-test. Transplant Proc 1995; 27: 3442-3.
8 Elin RJ. Magnesium metabolism in health and disease. Dis a Month 1988: 34: 161-218.
9 Durlach J. New trends in international magnesium research. Mag Res 1992; 5: 14-6.
10 Bley N. Dick K, Rob PM, Sack K. Follow-up measurement of serum, lymphocyte and red blood cell magnesium concentrations after renal transplantation and cyclosporin treatment with or without magnesium supplementation. In: Smetana R., eds. Advances in Magnesium Research London: John Libbey, 1997; 218-22.
11 Vormann J, Gunther T. Perras B. Rob PM. Magnesium metabolism in erythrocytes of patients with chronic renal failure and after renal transplantation. Eur J Clin Chem Clin Biochem 1994: 32: 901-4.
12 Elin RJ. Assessment of magnesium status. In: Itokawa Y, Durlach J, eds. Magnesium in Health and Disease. London: John Libbey, 1989; 137-46.
13 Fitzgerald MG, Fourman P. An experimental study of magnesium deficiency in man. Clin Sci 1956; 15: 635-47.
14 Thoren L. Magnesium deficiency in gastrointestinal fluid loss. Acta Chir Scand 1963; 306 (Suppl.): 1-65.
15 Dyckner T. Wester PO. Magnesium deficiency - guidelines for diagnosis and substitution therapy. Acta Med Scand 1982: 661 (Suppl.): 37-41.
16 Rasmussen HS, McNair P, Goransson L, Balslav S, Larsen OG, Aurup P. Magnesium deficiency in patients with ischemic heart disease with and without acute myocardial infarction uncovered by an intravenous loading test. Arch Intern Med 1988; 148: 329-32.
17 Cohen L, Kitzes H. Characterisation of the magnesium status of elderly people with congestive heart failure, hypertension, and diabetes mellitus by the magnesium-load test. Magnes Bull 1993; 15: 105-7.
18 Gullestad L, Dolva LO, Waage A, Falch D. Fagerthun H, Kjekshus J. Magnesium deficiency diagnosed by an intravenous loading test. Scand J Clin Lab Invest 1992; 52: 245-53.
19 Dick K, Bley N, Seyfert T, Dibbelt L. Rob PM. How to diagnose magnesium deficiency in outpatients. Trace Elem Electrolyte 1998; 15: 8-10.
20 Scoble JE. Freestone A, Varghese Z. Fernando ON, Sweny P, Moorhead J. Cyclosporine-induced renal magnesium leak in renal transplant patients. Nephrol Dial Transplant 1990: 5: 812-5.
21 Vormann J, Forster C, Zippel U, et al. Effects of magnesium deficiency on magnesium and calcium content in bone and cartilage in developing rats in correlation to chondrotoxicity. Calcif Tissue Int 1997: 61: 230-8.
22 Danielson BG, Johansson G, Ljunghall S. Magnesium metabolism in healthy subjects. Scand J Urol Nephrol 1979; 51 (Suppl.): 49-73.
23 Rude RK, Ryzen E. The Tm Mg and renal Mg threshold in normal man and in certain pathophysiologic conditions. Magnesium 1985; 5: 273-81.
Correspondence: Priv. Doz. Dr med. P.M. Rob, Ratzeburger Allee 160, D-23538 Lübeck, Germany (fax: 0049-451-5006334: email: email@example.com).
This page was first uploaded to The Magnesium Web Site on July 30, 2005