Original Article
Diagnosis of Magnesium-Induced Diarrhea
N Engl J Med 1991; 324:1012-1017April 11, 1991
- Abstract
- Abstract
Background.
There is no specific method of diagnosing magnesium-induced diarrhea. Therefore, the frequency and clinical importance of diarrhea caused by magnesium are unknown. The purposes of this study were to establish a method for diagnosing magnesiuminduced diarrhea and to apply it to patients with chronic diarrhea.
Methods.
We measured fecal output of soluble magnesium and fecal magnesium concentration in 19 normal subjects with formed stools (15 collection periods), with non-magnesium-induced diarrhea (36 collection periods), and with diarrhea induced by magnesium hydroxide alone (11 collection periods) or in combination with phenolphthalein (3 collection periods), and in 359 patients with chronic diarrhea.
Results.
The upper limits of fecal output of soluble magnesium and fecal magnesium concentration in normal subjects were 14.6 mmol per day and 45.2 mmol per liter, respectively. When normal subjects had diarrhea due to the ingestion of magnesium hydroxide alone or in combination with phenolphthalein, fecal magnesium output was always abnormally high. For each millimole increase in fecal magnesium output, fecal weight increased by approximately 7.3 g. The fecal magnesium concentration was very high when magnesium was the only cause of diarrhea but only moderately elevated when diarrhea was induced by magnesium hydroxide plus phenolphthalein. Biochemical and clinical evidence indicated that excessive ingestion of magnesium was an important cause of chronic diarrhea in 15 of the 359 patients with chronic diarrhea (4.2 percent), if not the only cause.
Conclusions.
Quantitative fecal analysis for soluble magnesium is an accurate method of diagnosing magnesium-induced diarrhea. Some patients with chronic diarrhea ingest excessive amounts of magnesium (in antacids or food supplements), and physicians may fail to discover this before embarking on an expensive and invasive diagnostic evaluation. (N Engl J Med 1991; 324:1012–7.)
Media in This Article
Figure 1
Fecal Output and Fecal Concentration of Soluble Magnesium (Mg) in Normal Subjects with Normal Stools or with Diarrhea Induced by Various Agents.Figure 2
Fecal Weight Plotted as a Function of Fecal Output of Soluble Magnesium in Normal Subjects with Diarrhea Induced by Three Different Doses of Magnesium Hydroxide (Solid Circles) (Linear Regression, y = 7.3x + 44.3).Article Activity
- Article
IT is well known that excessive ingestion of magnesium causes diarrhea. However, such diarrhea is not easily recognized because no diagnostic criteria for magnesium-induced diarrhea have been established. Although previous investigators have suggested upper limits of normal for the fecal magnesium concentration,1 2 3 4 5 these values have differed substantially. More importantly, the upper limit of normal for fecal magnesium output in people with diarrhea has not been determined.
The first purpose of our research was to establish a method of diagnosing magnesium-induced diarrhea. Because magnesium acts as an osmotic cathartic agent, magnesium-induced diarrhea should be associated with a high fecal concentration of soluble magnesium, a high fecal output of magnesium, or both. To establish the upper limits of normal for fecal magnesium concentration and output in people with diarrhea, diarrhea was induced in normal subjects by mechanisms that did not involve magnesium, and their stool was analyzed for magnesium content. (Some of these data have been published previously,6 but in a form that does not allow calculation of the upper limits of normal.) Diarrhea was then induced in normal subjects by giving them magnesium hydroxide to determine whether an abnormally high fecal magnesium content was an accurate indicator of magnesium-induced diarrhea. The second purpose of this study was to apply the criteria established for magnesium-induced diarrhea to a large group of patients with chronic diarrhea.
Methods
Normal Subjects
Nineteen normal subjects (18 men and 1 woman), ranging in age from 23 to 36 years, participated in a total of 65 experiments. They followed their customary diets and took no medications or vitamin or mineral supplements during the studies. These studies were approved by Baylor University Medical Center's Institutional Review Board for Human Protection; informed consent was obtained from all subjects.
We collected normal stool samples from 12 men during 15 separate 72-hour periods. We then induced osmotic diarrhea in these and other normal subjects by giving them a solution containing one of the following: 105 g of polyethylene glycol 3350 per liter, 75 g of sorbitol per liter, 50 or 70 g of lactulose per liter, 50 mmol of sodium sulfate per liter, or 27, 54, or 107 mmol of magnesium hydroxide per liter. Sodium chloride (48 mmol per liter), potassium chloride (5 mmol per liter), and sodium bicarbonate (17 mmol per liter) were added to each solution to prevent the development of electrolyte deficiency from fecal losses. On four consecutive days while at home, 14 men drank specified amounts of one of the solutions with each of three meals and a bedtime snack to provide a total daily dose of 53, 95, 189, or 252 g of polyethylene glycol 3350 (13 collection periods); 90 or 135 g of sorbitol (5 collection periods); 45, 95, or 125 g of lactulose (9 collection periods); 90 mmol of sodium sulfate (3 collection periods); or 48, 96, or 193 mmol of magnesium hydroxide (11 collection periods). We induced secretory diarrhea in six subjects (five men and one woman) by giving them 960 mg of phenolphthalein per day (Mission Pharmaceutical, San Antonio, Tex.) in four divided doses on four consecutive days.
To assess the effect of increased magnesium intake in the presence of diarrhea caused by an independent mechanism, diarrhea was induced in three men on two occasions by giving them 960 mg of phenolphthalein (in four divided doses) each day for four days. On the first occasion, they received phenolphthalein alone; on the second, they received phenolphthalein and 48 mmol of magnesium hydroxide each day.
During each of these studies, the subjects collected stool samples for 24 hours on each of the last three days of the four-day study. For subjects who participated in more than one study, the interval between studies was at least two weeks.
Patients with Chronic Diarrhea
For the past nine years at our center a determination of fecal magnesium content has been included in the evaluation of patients with diarrhea lasting one month or longer. During this period, 359 patients (137 male and 222 female) ranging in age from 8 to 77 years (mean, 47.6) were studied who had unformed stools during a 48- or 72-hour stool-collection period while following a normal diet. Patients who were referred for diarrhea but who had normal stools during the stool-collection period were excluded from this study.
Stool Collections and Analysis
Stools were quantitatively collected, weighed, and homogenized at low speed in a commercial blender (model 34BL22, Waring Products, New Hartford, Conn.). A weighed aliquot was centrifuged, and the supernatant used for the measurement of the soluble magnesium concentration by atomic absorption spectroscopy (coefficient of variation, 2.5 percent). A separate aliquot of the stool homogenate was lyophilized for measurement of the dry weight; the percentage of water in the stool was then calculated. The daily fecal output of soluble magnesium was calculated by multiplying the average daily fecal concentration of soluble magnesium by both the average daily stool weight for the collection period and the average daily water weight of stools. The two values correlated closely (r = 0.99, P<0.001). We will present here only the results based on stool weight, since stool lyophilization is not readily available in many laboratories. The results based on the fecal output of soluble magnesium calculated from stool water are available on request.
Results
Fecal Output of Soluble Magnesium in Normal Subjects
The mean (±SE) fecal output of soluble magnesium in normal subjects with no diarrhea was 5.1±0.6 mmol per day. When diarrhea was induced by various doses of polyethylene glycol, there was no concomitant increase in magnesium output as stool weight increased progressively up to 1600 g per day. The mean output of magnesium in subjects with diarrhea induced by polyethylene glycol, nonabsorbable carbohydrate, sodium sulfate, and phenolphthalein was 4.0±0.2, 6.1±0.8, 10.8±1.6, and 6.8±1.3 mmol per day, respectively. The output of magnesium in subjects with diarrhea induced by sodium sulfate and phenolphthalein was significantly higher than in subjects with diarrhea caused by polyethylene glycol (P = 0.0004 for sodium sulfate, P = 0.04 for phenolphthalein).
Figure 1Figure 1
Fecal Output and Fecal Concentration of Soluble Magnesium (Mg) in Normal Subjects with Normal Stools or with Diarrhea Induced by Various Agents.A shows the fecal output of soluble magnesium in individual subjects. The mean (±SE) output for the 51 collections from normal subjects who had not been given magnesium (15 normal-stool collections and 36 unformed-stool collections) was 5.6±0.4 mmol per day, and the standard deviation was 3.0 mmol per day. To calculate the upper limit of normal we used 3 rather than 2 SD in order to enhance specificity (reduce the number of false positive results). According to this definition, the upper limit was 14.6 mmol per day.The values for fecal output of soluble magnesium for individual subjects with magnesium-induced diarrhea were all abnormally high (Fig. 1A); the magnitude of the elevation was directly proportional to the dose of magnesium hydroxide ingested. The output during the ingestion of magnesium hydroxide and phenolphthalein was approximately the same as that observed when magnesium hydroxide was ingested alone.
Fecal Concentration of Soluble Magnesium in Normal Subjects
The fecal concentrations of soluble magnesium in individual subjects are shown in Figure 1B. As we did for magnesium output, we calculated the upper limit of normal for the fecal concentration of soluble magnesium from values obtained in subjects with formed stools and those with liquid stools due to mechanisms that did not involve an excessive intake of magnesium. The mean (±SE) was 14.9±1.4 mmol per liter, the standard deviation was 10.1 mmol per liter, and the upper limit of normal (mean +3 SD) was 45.2 mmol per liter.
The fecal concentration of soluble magnesium in subjects given magnesium hydroxide averaged 119, 128, and 134 mmol per liter with the low, medium, and high doses of magnesium hydroxide, respectively. By contrast, the fecal concentration of magnesium in three subjects given a combination of magnesium hydroxide and phenolphthalein averaged approximately 50 mmol per liter.
Other Observations in Normal Subjects
When diarrhea was induced in five normal subjects with different doses of magnesium hydroxide (11 collection periods), the fecal output of soluble magnesium was closely correlated with the amount of elemental magnesium consumed as magnesium hydroxide (r = 0.98, P<0.001). Moreover, fecal weight was closely correlated with the fecal magnesium output (r = 0.99, P<0.001) (Fig. 2Figure 2
Fecal Weight Plotted as a Function of Fecal Output of Soluble Magnesium in Normal Subjects with Diarrhea Induced by Three Different Doses of Magnesium Hydroxide (Solid Circles) (Linear Regression, y = 7.3x + 44.3).). The slope of the regression line in Figure 2 indicates that when magnesium is the sole cause of diarrhea, an incremental output of 1 mmol of soluble magnesium results in a 7.3-g increment in fecal weight.In three subjects who ingested phenolphthalein together with magnesium hydroxide, the average increment in fecal weight due to a 1-mmol increment in fecal magnesium output was 8.1 g — a value similar to the increment of 7.3 g per millimole observed when magnesium was the sole cause of diarrhea. In these three subjects, fecal weights were 500 to 600 g higher than would be predicted from their soluble magnesium outputs (Fig. 2).
Patients with Chronic Diarrhea
Twenty-one of the 359 patients with chronic diarrhea (5.8 percent) had an abnormally high fecal output or concentration of soluble magnesium. Eleven of these patients had both a high fecal output and a high fecal concentration of magnesium. In nine the output was high but the concentration was within normal limits, and in one patient the concentration was high but the output was normal.
In 15 of these 21 patients (4.2 percent of the total population), an unsuspected nondietary source of magnesium was established. This was confirmed in 14 patients by simple questioning; the sources are shown in Table 1Table 1
Characteristics of Patients with Fecal Magnesium Concentrations or Outputs above the Upper Limits of Normal.. The response of 13 of these patients to a reduction in magnesium intake was assessed by history; all 13 reported an improvement or cessation of their diarrhea. The mother of Patient 15, an eight-year-old girl, denied having given the child magnesium hydroxide, but it was found during a search of the child's hospital room. The child's life-threatening diarrhea resolved promptly and permanently after a court-ordered separation from the mother. This patient is also of special interest because, although her fecal magnesium output was extremely high (76 mmol per day), her fecal magnesium concentration (31 mmol per liter) was well within the normal range and her fecal weight of 2445 g per day was well above that expected on the basis of magnesium ingestion alone. Thus, this patient was probably receiving a second (but unidentified) laxative at the same time.The data for the six patients in whom extradietary magnesium intake could not be established are also summarized in Table 1. In most instances fecal magnesium output was only slightly elevated, and five of the patients had steatorrhea, suggesting that intestinal malabsorption of magnesium in the presence of a normal dietary intake may have been responsible for their high fecal output of magnesium.
Figure 3Figure 3
Fecal Weight Plotted as a Function of Fecal Output of Soluble Magnesium in 21 Patients with High Fecal Magnesium Values. shows the daily fecal weight plotted as a function of the fecal output of soluble magnesium in the 21 patients who had elevated fecal magnesium values. The line shown is the linear regression line from Figure 2 — that is, the relation between the fecal weight and fecal magnesium output in the normal subjects in whom magnesium was the only cause of diarrhea. If fecal magnesium acted in these patients as it did in normal subjects (i.e., an increment of 1 mmol of soluble magnesium on average increased fecal weight by about 7.3 g, whether or not magnesium was the sole cause of the diarrhea), then the fecal weight that would be expected on the basis of the fecal output of soluble magnesium could be estimated. The results for the patients in whom magnesium was the only cause of diarrhea should lie near the regression line (Fig. 3); this was true for five of the patients (Patients 1 to 5 in Table 1). Nine other patients (Patients 6 to 14), all with confirmed magnesium ingestion from a variety of sources (Table 1), had stool weights about twice that expected on the basis of their fecal magnesium output, suggesting that magnesium was an important cause of diarrhea but not the only cause. In the remaining patients (Patients 15 to 21), including those with steatorrhea but no evidence of a high magnesium intake, the stool weight was 5 to 35 times higher than could be explained on the basis of the fecal magnesium values, and other causes were therefore mainly responsible for diarrhea.Discussion
To establish criteria for the diagnosis of magnesium-induced diarrhea on the basis of fecal analysis, we induced diarrhea in normal subjects (on normal diets) with several agents other than magnesium. Ingestion of polyethylene glycol (an inert osmotic agent) increased fecal weight to as much as 1600 g per day but had no effect on the output of soluble magnesium. Because the fecal output of magnesium remained relatively constant, the soluble magnesium concentration fell in a curvilinear fashion as diarrhea developed and worsened. Phenolphthalein-induced diarrhea was associated with a significantly higher fecal magnesium output than diarrhea induced by polyethylene glycol, perhaps because of a phenolphthalein-related disturbance of mucosal-cell absorptive and secretory processes.7 8 9 The magnesium output during sodium sulfate—induced diarrhea was almost three times that caused by polyethylene glycol — presumably an effect of the poorly absorbed sulfate anion.
On the basis of the results in normal subjects, both when they had normal formed stools and when diarrhea was induced by mechanisms other than increased magnesium intake, the upper limits of normal for the fecal output of soluble magnesium and the fecal magnesium concentration were 14.6 mmol per day and 45.2 mmol per liter, respectively. The results in all subjects with induced diarrhea, except that due to magnesium, were well within these limits, whereas in normal subjects with diarrhea induced solely by excessive intake of magnesium, the results exceeded the upper limits of normal. When diarrhea was induced by a combination of phenolphthalein and magnesium, the fecal output of magnesium was the same as when magnesium was the only cause of diarrhea; the fecal magnesium concentration, however, was only slightly higher than the upper limit of normal. We conclude that a fecal output of soluble magnesium that exceeds 14.6 mmol per day or a fecal magnesium concentration that exceeds 45.2 mmol per liter (or both) is abnormal and suggests that magnesium has a role in the pathogenesis of diarrhea.
These studies in normal subjects provide several new insights into the pathophysiologic processes of magnesium-induced diarrhea. First, the fecal output of soluble magnesium is closely correlated with the amount of magnesium ingested as magnesium hydroxide, suggesting that most of the nondietary magnesium is unabsorbed and soluble in the water phase of stool. Second, daily fecal weight is closely correlated with fecal output of soluble magnesium, so that an increment of 1 mmol of soluble magnesium increases the fecal weight by 7.3 g, mainly because of an increase in the amount of water in the stool. Each millimole of fecal magnesium is dissolved in 7.3 g (0.0073 liter) of stool water, so that the fecal concentration of soluble magnesium in magnesium-induced diarrhea rapidly exceeds normal levels and ultimately approaches 137 mmol per liter. Finally, in the context of a high magnesium output, a fecal concentration of soluble magnesium above 100 mmol per liter suggests that magnesium is the only cause of diarrhea (Fig. 1B). On the other hand, when magnesium output is high but the fecal concentration of soluble magnesium is less than 100 mmol per liter, magnesium is not the only cause of diarrhea. The degree to which excess fecal magnesium is the cause of an individual case of diarrhea can be estimated by plotting fecal weight as a function of fecal magnesium output, as shown in Figure 3.
Another purpose of our research was to apply the criteria for magnesium-induced diarrhea that we had established to 359 patients with chronic diarrhea. We found that 21 of these patients (5.8 percent) had abnormally high levels of fecal magnesium output or abnormally high fecal magnesium concentrations, or both, in most instances high enough to contribute substantially to the patient's diarrhea. In the majority of patients this could be attributed to the ingestion of nondietary magnesium from various sources, suspected only after a high fecal magnesium output or concentration was discovered. The patients were unaware of the diarrhea-producing potential of these products, and they had not been asked about the use of magnesium-containing products by their physicians. One patient, an eight-year-old girl, had been given milk of magnesia surreptitiously by her mother. In several other patients, malabsorption of magnesium in the presence of a normal dietary intake was the most likely explanation for the elevation of fecal magnesium concentrations. Because the normal dietary intake of magnesium is relatively small, malabsorption of magnesium makes only a minor contribution to the diarrhea of the malabsorption syndrome.
We reached four main conclusions. First, analysis of quantitatively collected stool specimens for soluble magnesium is an accurate method of diagnosing magnesium-induced diarrhea; quantitative analysis of fecal magnesium output is a necessary component of this method because if only the fecal magnesium concentration is measured, the diagnosis may be missed when more than one cause of diarrhea is present. In addition, the method allows assessment of the degree to which fecal magnesium is responsible for the excessive excretion of stool water. Second, patients with chronic diarrhea may ingest excessive amounts of magnesium (in antacids or food supplements) and be unaware that this can cause or exacerbate diarrhea. Third, even experienced physicians may fail to exclude excessive magnesium intake as a cause of diarrhea before embarking on an expensive and invasive diagnostic evaluation. And finally, magnesium-induced chronic diarrhea often remains unrecognized until fecal analysis for magnesium is carried out.
Supported by a grant (5–R01–DK37172–05) from the National Institute of Diabetes and Digestive and Kidney Diseases and by the Southwest Digestive Disease Foundation. Dr. Fine is the Reba Perry Powell Fellow in Gastrointestinal Research.
We are indebted to Sharon Michael for assistance in the preparation of the manuscript and to Jack Porter and Diana Santa Ana for expert technical assistance.
Source Information
From the Department of Internal Medicine, Baylor University Medical Center, 3500 Gaston Ave., Dallas, TX 75246, where reprint requests should be addressed to Dr. Fordtran.
- References
References
1
Wrong
Web of Science | Medline2
Bjork
Web of Science3
Morris
Web of Science | Medline4
Krejs GJ, Hendler RS, Fordtran JS. Diagnostic and pathophysiologic studies in patients with chronic diarrhea. In: Field M, Fordtran JS, Schultz SG, eds. Secretory diarrhea. Bethesda. Md.: American Physiological Society, 1980: 141–51.
5
Saunders
Medline6
Hammer
CrossRef | Web of Science | Medline7
French
CrossRef | Web of Science | Medline8
Binder
Web of Science | Medline9
Gaginella
CrossRef | Web of Science | Medline
- Citing Articles (19)
Citing Articles
1
Katarzyna Wyskida, Joanna Witkowicz, Jerzy Chudek, Andrzej Więcek. (2011) Daily Magnesium Intake and Hypermagnesemia in Hemodialysis Patients With Chronic Kidney Disease. Journal of Renal Nutrition
CrossRef2
James L. Roerig, Kristine J. Steffen, James E. Mitchell, Christie Zunker. (2010) Laxative Abuse. Drugs 70:12, 1487-1503
CrossRef3
Lawrence R. Schiller. (2009) Diarrhea and Malabsorption in the Elderly. Gastroenterology Clinics of North America 38:3, 481-502
CrossRef4
K. Holtenius, C. Kronqvist, E. Briland, R. Spörndly. (2008) Magnesium Absorption by Lactating Dairy Cows on a Grass Silage-Based Diet Supplied with Different Potassium and Magnesium Levels. Journal of Dairy Science 91:2, 743-748
CrossRef5
Rajesh N. Keswani, Jenny Sauk, Sunanda V. Kane. (2006) Factitious Diarrhea Masquerading as Refractory Celiac Disease. Southern Medical Journal 99:3, 293-295
CrossRef6
J.G. Gobert, L. Barbot, N. Kapel. (2004) Élément de physiopathologie digestive et coprologie fonctionnelle. Annales Pharmaceutiques Françaises 62:6, 367-370
CrossRef7
A. Villanueva Rodríguez. (2004) Síndrome diarreico en el adulto. Medicine - Programa de Formación Médica Continuada Acreditado 9:3, 185-192
CrossRef8
L. R. Schiller. (2001) The therapy of constipation. Alimentary Pharmacology and Therapeutics 15:6, 749-763
CrossRef9
Andrew Duncan, John A. H. Forrest. (2001) Surreptitious abuse of magnesium laxatives as a cause of chronic diarrhoea. European Journal of Gastroenterology & Hepatology 13:5, 599-601
CrossRef10
Lawrence R. Schiller. (2000) DIARRHEA. Medical Clinics of North America 84:5, 1259-1274
CrossRef11
Lissy de Ridder, J. Hans Hoekstra. (2000) Manifestations of Munchausen Syndrome by Proxy in Pediatric Gastroenterology. Journal of Pediatric Gastroenterology and Nutrition 31:2, 208-211
CrossRef12
Lawrence R. Schiller. (1999) Clinical Pharmacology and Use of Laxatives and Lavage Solutions. Journal of Clinical Gastroenterology 28:1, 11-18
CrossRef13
Topazian, MarkBinder, Henry J.. (1994) Factitious Diarrhea Detected by Measurement of Stool Osmolality. New England Journal of Medicine 330:20, 1418-1419
Full Text14
J Carlson, P Fernlund, S-A Ivarsson, I Jakobsson, J Neiderud, KO Nilsson, M Svensson, U Swanstein. (1994) Munchausen syndrome by proxy: an unexpected cause of severe chronic diarrhoea in a child. Acta Paediatrica 83:1, 119-121
CrossRef15
G. Bleichner, O. Thomas, J.P. Sollet. (1993) Diarrhea in intensive care: diagnosis and treatment. International Journal of Antimicrobial Agents 3:1, 33-48
CrossRef16
Sherry L. Perkins, John F. Livesey. (1993) A rapid high-performance thin-layer chromatographic urine screen for laxative abuse. Clinical Biochemistry 26:3, 179-181
CrossRef17
ROGER BOUILLON, LUC VERRESEN, FILIP STAELS, MARIE BEX, PAUL DE VOS, MICHEL DE ROO. (1993) The Measurement of Fecal Thyroxine in the Diagnosis of Thyrotoxicosis Factitia. Thyroid 3:2, 101-103
CrossRef18
Afzalpurkar, Rekha G., Schiller, Lawrence R., Little, Katherine H., Santangelo, William C., Fordtran, John S., . (1992) The Self-Limited Nature of Chronic Idiopathic Diarrhea. New England Journal of Medicine 327:26, 1849-1852
Full Text19
Donowitz, Mark, . (1991) Magnesium-Induced Diarrhea and New Insights into the Pathobiology of Diarrhea. New England Journal of Medicine 324:15, 1059-1060
Full Text
