Because magnesium has so many critical functions, the body has mechanisms to assure that magnesium is readily available. During low intakes of magnesium, the percent absorbed from the diet is increased, the amount excreted in the urine is decreased, and body reserves, mainly in bone, are used. When the dietary intake of magnesium is adequate, the opposite occurs with magnesium increasing in urine and body reserves while the percentage absorbed from the diet is decreased [1, 2].
The response of the body to maintain magnesium homeostasis when changes in dietary intakes occur has made it difficult to establish status indicators and set dietary requirements for magnesium. In the past 15 years, balance data, much from controlled metabolic unit studies, and meta-analyses of magnesium experiments have provided some guidance for setting magnesium requirements and determining the usefulness of urinary and serum magnesium as indicators of magnesium status.
Based on the Dietary Reference Intakes (DRIs) set throughout the world, magnesium deficiency commonly occurs. Yet, pathological conditions attributed specifically to dietary magnesium deficiency alone (not taking medications or having disorders that inhibit absorption or induce excretion of magnesium) often are considered rare. In the United States, a possible reason for these differing conclusions is that unreliable balance data were used to set the DRIs for magnesium in 1997. The Estimated Average Requirement (EAR) and Recommended Dietary Allowance (RDA) for adult women were set at 255-265 mg and 310-320 mg/day, respectively. The EAR and RDA, respectively, for adult men were set at 330-350 mg and 410-420 mg/day . These DRIs were based on highly variable balance data from 16 men and 18 women on self-selected diets that decreased in magnesium during the balance periods , which could have influenced balance values. Some subjects on magnesium intakes less than 258 mg/day had equilibrium or positive magnesium balance, while some subjects with intakes greater than 299 mg/day had a negative balance. The DRIs based on these balance findings have been questioned, and based on the subsequent balance data from a large number of individuals with controlled dietary intakes of magnesium, may be too high.
Balance data were determined in 27 different metabolic unit studies of 93 men and 150 women in which dietary magnesium ranged from 84 to 598 mg/day . The ages of the subjects ranged from 19 to 77 years, and weight ranged from 46 to 136 kg. In these studies, the metabolic unit provided a common environment for strict control of food consumption, weight, exercise, and fecal and urine collection. The participants consumed only food and beverages provided by the dietary staff and were chaperoned on all outings from the metabolic unit to ensure compliance with study protocols. The methods used for the balance determinations were similar for all experiments. Duplicate diets of each volunteer in 14 studies, and one representative diet supplying 2000 kcal (8.4 MJ) in each of the other 13 studies were prepared daily for analysis. Urine and feces were collected completely in plastic containers and bags, respectively. The magnesium content of 6-7 day composites of diets, urine, and feces were determined by flame atomic absorption spectrometry (studies 1-6) or by inductively coupled plasma emission spectroscopy (studies 7-27). Balances were calculated as the difference between dietary intake and urinary plus fecal excretion of magnesium. The balance calculations did not include surface, phlebotomy, menstrual, or seminal losses; these losses probably would have lowered balance by another 3-6 mg/day.
The balance data from these 27 studies revealed that neutral magnesium balance occurred at an magnesium intake of 165 mg/day with a 95% prediction interval of 113 and 237 mg/day. Using the upper 95% value of 237 mg/day and considering that 98% is the upper interval level used for setting the RDAs, would result in an RDA of 245 mg/day. Considering surface, phlebotomy, menstrual, and seminal losses would increase the RDA to 250 mgday. Using the value of 165 mg/day and the same manipulations as done for the RDA would yield an EAR of 175 mg/day.
The suggested RDA of 250 mg/day is supported by numerous studies showing an association between dietary magnesium and chronic inflammatory stress evidenced by elevated C-reactive protein ; cardiovascular disease risk , events , and mortality ; and the metabolic syndrome [10, 11]. In the studies described by these groups, the most significant differences occurred when the mean lowest intake was ≤250 mg/day.
The balance data also found a neutral magnesium balance per kg of body weight of 2.36 mg/day with a 95% prediction interval of 1.58 and 3.38 mg/day . These data indicate that individuals weighing more or less than 70 kg should have RDAs higher or lower, respectively, than 250 mg/day. For example, the suggested RDA for individuals weighing 100 kg would be 355 mg/day and for individuals weighing only 50 kg, 175 mg/day. A dietary requirement for magnesium based on body weight may be the reason DRIs are usually set higher for men than women; they are more likely to weigh more. A higher requirement for magnesium also may be the reason for reports that a low magnesium status occurs more often in obese than non-obese individuals [12-14]. It also should be noted that the balance data used for the suggested DRIs were obtained from generally healthy individuals under optimal dietary and environmental conditions. Magnesium requirements may be increased by factors such as increased oxidative and inflammatory stress , foods (e.g., high in fiber/phytate) and medications (e.g., proton pump inhibitors) that inhibit absorption, and excessive alcohol intake or medications (e.g., diuretics) that increase excretion [1, 15].
Other reports support the suggested DRIs above. Balance data from German women and men indicated that normative magnesium requirements were less than 200 mg/day for women and 250 mg/day for men, and less than 3.0 mg/kg body weight/day . A study with female adolescents weighing about 55 kg indicated that a magnesium intake of 2.94 mg/kg body weight/day or 150 mg/day was inadequate, and balance data indicated that 4.25 mg/kg body weight/day or 232 mg/day was adequate .
One report does not support the suggested DRIs above  Combined data from 13 studies involving131 young female subjects found an estimated equilibrated dietary magnesium intake of 4.18 mg/kg body weight/day. This contrasting finding might have occurred because of the short durations of the balance studies (8-12 days). The metabolic unit studies found that extended balance periods were need to get acceptable balance data because of daily variation in the excretion of magnesium.
Urine as a status indicator
Recently, urinary magnesium excretion was associated with the risk of cardiovascular disease [19, 20] and hypertension [20, 21]. In one study, 7664 adults were divided into quintiles base on urinary magnesium excretion . The lowest sex-specific quintile [<71 mg (2.93 mmol)/day for men and <60 mg (2.45 mmol)] for women had an increased risk for fatal and nonfatal ischemic heart disease compared to the upper four quintiles. In the other studies involving 12,335 middle-aged men and women  and 5511 participants aged 28-75 years , an inverse relation was found between 24-hour urinary magnesium and the prevalence or development of hypertension. These findings suggest that urinary magnesium might be an indicator of magnesium status. This suggestion is supported by the urinary excretion data obtained from the 27 metabolic unit studies described above. These data show that when intakes were 200 mg/day or lower urinary magnesium excretion generally ranged from 40 to 80 mg (1.65 to 3.29 mmol)/day. When dietary magnesium intakes were greater than 250 mg/day, urinary magnesium excretion general ranged from 80 to 160 mg (3.29 to 6.58 mmol)/day [22, 23]. Four depletion-repletion experiments, using methods described for the 27 metabolic unit studies described above, support the indicated relation between urinary and dietary magnesium . These experiments found that magnesium urinary excretion was between 40 and 80 mg (1.65 and 2.88 mmol)/day with dietary intakes less than the suggested EAR of 175 mg/day. With intakes above 300 mg/day, urinary magnesium excretion ranged from 100 to 140 mg (4.11 to 5.76 mmol)/day.
These experiments also found that urinary magnesium changed rapidly with a change in dietary magnesium . An example of this change is shown in figure 1, which give results from a double-blind crossover experiment [22, 24]. After an equilibration period of 10 days with a mean dietary magnesium intake of 299 mg/day (dietary copper was 1 mg/day), postmenopausal women that consumed a mean intake of 390 mg/day for 81 days had a mean daily urinary excretion of 120 -130 mg (4.93-5.35 mmol)/day. When the women were switched to a diet supplying a mean of 120 mg/day, mean urinary magnesium excretion decreased to 63 mg (2.59 mmol)/day within 6 days and remained near this level for the remaining 75 days of magnesium depletion. Women that were placed on a diet providing a mean intake of 116 mg/day for 81 days after the equilibrium period exhibited a decrease in urinary magnesium to 58 mg (2.39 mmol)/day within 6 days and remained near that level the last 75 days of the depletion period. When these women were switched to a diet supplying a mean intake of 426 mg/day, urinary magnesium excretion increased to 132 mg (5.43 mmol)/day within 6 days and remained near that level for the remaining 75 days of high magnesium intake. These findings indicate a single determination or short-term 24-hour urinary magnesium determinations is not useful as an indicator of current magnesium status because magnesium excretion may be low while an individual still has an adequate status, and vice versa. However, urinary magnesium can be an indicator of magnesium intakes in population studies, which would give an estimate of the percent of individuals not consuming the RDA or EAR for magnesium. These individuals might be at risk for pathology associated with a low magnesium status.
Serum as a status indicator
Currently, a magnesium concentration below 0.75 mmol/L or 1.82 mg/dL often is used as an indication of magnesium deficiency. Using this value to indicate magnesium deficiency is supported by numerous studies showing an association between magnesium deficiency and chronic disease [7, 8, 25]. However, changes in serum magnesium shown graphically in these reports [7, 8, 25] indicate that many individuals with concentrations between 0.75 mmol(1.82 mg/dL) and 0.80 mmol/L (1.94 mg/dL) also have an increased chronic disease risk. In addition, four depletion/repletion metabolic unit studies  found that individuals with serum magnesium concentrations near 0.85 mmol/L (2.06 mg/dL) may be magnesium-deficient because they exhibited physiological and biochemical changes that responded to magnesium supplementation. An example of the findings from these studies is shown in figure 2. In the experiment where the postmenopausal women were on an equilibration period of 10 days with a mean dietary magnesium intake of 299 mg/day (dietary copper was 1 mg/day), mean serum magnesium concentration did not vary much from 0.84 mmol/L (2.05 mg/dL) when they consumed a mean intake of 390 mg/day for 81 days. When the women were switched to a diet supplying a mean of 120 mg/day, it appeared that the concentration decreased to 0.80 mmol/L (1.95 mg/dL) 60 days later, however, the concentration subsequently returned to 0.84-0.87 mmol/L (2.05-2.12 mg/dL). Other metabolic unit studies  also found that consuming a magnesium-deficient diet for 72 to 92 days did not markedly decrease serum magnesium concentrations. The lack of response indicates that it takes an extended period of time for serum magnesium to decrease when individuals begin with apparent ample magnesium stores or status. This finding is consistent with the conclusion of Elin  that changes in serum magnesium concentration to a new equilibrium normally occur very slowly over a period of months to years.
Women that were placed on a diet providing a mean intake of 116 mg/day for 81 days after the equilibrium period exhibited a slight increase in serum magnesium concentration during the first 40 days before decreasing below 0.82 mmol/L (2.0 mg/dL) and then increasing so the final concentration was 0.85 mmol/L (2.06 mg/dL) (figure 2). When these women were switched to a diet supplying a mean intake of 426 mg/day, serum magnesium increased to 0.95 mmo/L (2.3 mg/dL) and remained at near this level for 30 days before decreasing to 0.92 mmol (2.2 mg/dL) at the end of the repletion period. Although mean magnesium concentrations never decreased below a mean 0.80 mmol/L (1.95 mg/dL) during the 81 day-depletion periods in this study, some women apparently were magnesium deficient . Three women consuming the magnesium-deficient diet providing copper at 1 mg/day exhibited heart rhythm changes that responded to magnesium supplementation but not copper supplementation .
Recently, a meta-analysis study of 48 magnesium supplementation trials (n = 2131) found a dose-response with circulating magnesium . When baseline concentrations increased the increase in circulating magnesium in response to supplementation decreased. With baseline concentrations ≥0.87 mmol (2.12 mg/dL), magnesium supplementation did not result in significant changes in circulating magnesium concentrations. The findings of the meta-analysis and those from the metabolic unit depletion/repletion experiments  provide evidence that using a serum magnesium concentration >0.75 mmol/L (1.82 mg/dL) to assess an individual as not magnesium-deficient is faulty. As indicated by Elin  and by a recent evaluation of the reference range , many individuals with serum magnesium concentrations between 0.75 and 0.85 mmol/L (1.82 and 2.06 mg/dL) could be magnesium-deficient.
The determination of the occurrence of magnesium deficiency and whether it is of major nutritional concern for health and well-being has been hampered by using questionable DRIs and a faulty reference range for serum magnesium. Recent balance date indicate that the magnesium EAR and RDA for 70-kg healthy individuals may be 175 and 250 mg/day, respectively, and these decrease or increase based on body weight. Using the current reference range for serum magnesium of 0.750 to 0.955 mmol/L (1.82 to 2.32 mg/dL) for an indication of an adequate magnesium status needs revision because it can lead to faulty conclusions. Recent reports indicate that individuals with serum magnesium concentrations >0.75 mmol/L (1.82 mg/dL), or as high 0.85 mmol/L (2.06 mg/dL), could be magnesium-deficient. Thus, to assess magnesium status of an individual with a serum magnesium concentration between 0.75 and 0.85 mmol/L (1.82 and 2.06 mg/dL) requires additional measures of status. A urinary excretion of < 80 mg (3.29 mmol)/day and/or dietary intake history showing a magnesium intake of <250 mg/day would support the presence of magnesium deficiency, especially if there is the presence of a chronic disease associated with inflammatory stress.
Financial support: none. Conflict of interest: none.
* Presented at The XIV International Magnesium Symposium, Magnesium and Health, Rome, Italy, June 23-24, 2016.