Magnesium Concentration in Dialysate

In this issue of the Clinical Journal of the American Society of Nephrology, Bressendorff et al. report a single-center, randomized, controlled, double-blind trial in patients on maintenance hemodialysis (1). A total of 59 patients were randomized to either standard dialysate magnesium (0.5 mmol/L; 1 mEq/L) or to a high-dialysate magnesium (1 mmol/L; 2 mEq/L) for 4 weeks and then followed on standard dialysate magnesium for another 2 weeks. The primary end point was the change in a global serum calcification assay, T50, which measures the functional performance of the mineral buffering system in blood (see below). At the end of the initial 4-week period, T50 time had increased by 22% in the high-magnesium group but remained stable in the control group. This increase in T50 indicates a higher potency of the serum to inhibit calcification. The study by Bressendorff et al. (1) extends and confirms prior experiences of the same group in patients with CKD stages 3 and 4, where 8 weeks of oral magnesium supplementation at 30 mmol/d also increased T50 times by about 15% (2).

The study Bressendorff et al. (1) should be analyzed from two different angles: first, what is the relevance of a 22% increase in T50? Second, how much does this article contribute to the discussion on what constitutes an optimal level of dialysate magnesium?

Consider the serum calcification propensity test (T50) somewhat analogous to the complement CH50 assay. In the CH50 assay, an integrative biologic test is used to assess the functional intactness of the classic complement system, i.e., the ability of serum to lyse IgG-coated sheep red cells. The T50 test was developed to comprehensively measure the capacity of serum to inhibit calcification or, more specifically, to measure how long it takes for early, primary calciprotein particles to convert to secondary calciprotein particles (3). Early forms of calciprotein particles, including primary calciprotein particles, are likely physiologic transporters for bulk minerals in the bone compartment or, in patients with CKD, excess circulating mineral complexes that are rapidly cleared from the circulation, whereas secondary calciprotein particles have been used to induce vascular smooth muscle calcification (4). The longer the T50 time, the more potently that serum inhibits the conversion of primary to secondary calciprotein particles. A number of clinical studies have demonstrated that longer T50 times are associated with fewer cardiovascular events and mortality in patients with CKD, on dialysis or with a kidney allograft (3). At least in vitro, T50 times depend on concentrations and the interplay of various calcification-inhibiting factors, including fetuin-A and albumin, and ions such as calcium, phosphate, magnesium, and pyrophosphate (3). The study by Bressendorff et al. is important because it prospectively and in vivo establishes that therapeutic manipulation of magnesium levels in blood can affect T50 times just as in vitro. In a similar approach, 3 months of acetate-free, citrate-acidified bicarbonate dialysis also led to a 17% increase in T50 time, which was related to increases in serum albumin after the switch (5). Whether these short-term observations translate into any long-term benefits is the obvious question, and this is where matters become more complex.

Raising dialysate magnesium levels is a simple and effective way of increasing serum magnesium levels in patients on hemodialysis. Indeed, although the trial of Bressendorff et al. only lasted 4 weeks, serum magnesium levels rose from 0.99 to 1.40 mmol/L (2.4–3.4 mg/dl), i.e., the upper limit of the normal range (0.75–1.50 mmol/L; 1.8–3.6 mg/dl). Whether supplementing magnesium will affect extraosseous cardiovascular calcifications is currently being evaluated by Bressendorff et al. in the effect of magnesium supplementation on vascular calcification in chronic kidney disease (MAGiCAL-CKD) trial (Clinicaltrials.gov: NCT02542319), which will assess progression of coronary artery calcification over 1 year in patients with CKD stages 3b and 4, randomized to slow-release magnesium hydroxide or placebo. This latter trial will hopefully provide important information on magnesium’s anticalcification effects, and it will extend data of several short-term, small studies in patients with CKD demonstrating slower or lesser cardiovascular calcification with magnesium-containing phosphate binders compared with calcium-containing binders (6). At least theoretically, it is even conceivable that magnesium supplementation might promote calcification by suppressing parathyroid hormone (PTH) release, thus contributing to adynamic bone disease, a known risk factor for cardiovascular calcifications (7). Indeed, over the 4-week intervention in Bressendorff et al.’s trial, median intact PTH levels fell from 328 to 238 pg/ml (−27.5%) in the high-magnesium dialysate group. Besides suppressing PTH release, magnesium can also directly affect bone mineralization via interfering with calcium phosphate crystal formation just as in the vasculature. This may explain why, in an Italian trial published 30 years ago, lowering of dialysate magnesium from 0.5 to 0.25 mmol/L (1.0–0.5 mEq/L) significantly improved signs of osteomalacia in bone biopsies in normo-magnesemic patients on hemodialysis (8). Although measurements of bone mineral density and biochemical parameters of bone turnover will be part of the MAGiCAL-CKD trial, bone biopsies are not planned. We will therefore not know whether mineralization defects, in particular osteomalacia, increase upon magnesium supplementation.

Besides calcification and bone histology, how likely is it that increasing serum magnesium in patients with CKD or on dialysis, be it via the dialysate or oral, will affect morbidity and mortality? Low serum magnesium levels are common in many patients on dialysis and 5% are overtly hypomagnesemic in those dialyzed against a 0.5 mmol/L (1.0 mEq/L) bath (7). This is in part because of nutritional restrictions, removal of magnesium during dialysis if its dialysate concentration is below 1 mmol/L, low levels of active vitamin D, and some medications that have modest effects on serum magnesium levels (particularly proton pump inhibitors) (7,9). However, a caveat in all studies on serum magnesium is that usually total magnesium is measured, of which only 60%–70% is freely diffusible and not bound to proteins or complexed with anions. Keeping this caveat in mind, a large number of association studies in patients with CKD and other patient groups have associated low serum magnesium levels with poor clinical outcomes, cardiovascular disease, sudden cardiac death, metabolic syndrome, diabetes, hypertension, fatigue, and depression, all of which are particularly common in patients with CKD (10).

In the context of the above discussion, it is notable that in the trial of Bressendorff et al. (1) there were seven vascular events in the high-magnesium group, ranging from ischemic stroke to problems with dialysis access, versus one vascular event in the low-magnesium group, and all this despite a baseline imbalance in diabetes prevalence (41% diabetes in the low-magnesium group versus 21% diabetes in the high-magnesium group), which would have placed the low-magnesium group at a higher cardiovascular risk. Whether the higher vascular event rate was a chance finding or a real effect of higher magnesium dialysate remains to be determined. In a recent retrospective case-control study, 25 patients on a dialysate magnesium of 0.75 mmol/L were matched with 50 patients on a dialysate magnesium of 0.50 mmol/L (11). During a 3-year follow-up, 20% of the patients died in the high-magnesium group versus 36% in the low-magnesium group. Cox proportional hazards regression yielded a 65% risk reduction for all-cause mortality and significantly lower cardiovascular mortality with an increased dialysate magnesium level in patients on maintenance dialysis.

To conclude, what is the relevance of a 22% increase in T50 and how much does the study by Bressendorff et al. (1) contribute to the discussion of what constitutes an optimal level of dialysate magnesium? Given its short-term nature and biochemical end point, the study certainly adds only a small piece to the puzzle. However, it is an important trial because it is the first to confirm that a simple intervention, namely increasing dialysate magnesium to 1 mmol/L, can affect serum’s propensity to inhibit calcification. Whether this will ultimately reduce hard clinical end points and whether it is safe in the long run remains to be determined. For now, in the majority of patients on hemodialysis, the commonly used dialysate magnesium concentration of 0.5 (or rarely, 0.75 mmol/L) still appears acceptable, pending further data (6).

Disclosures

J.F. reports having received advisor/consultant fees from Amgen, Astellas, Bayer, Chugai, Medice, and Vifor, and speaker fees from Amgen and Fresenius Medical Care.