Hypomagnesaemia, the kidney and the vessels

The morbidity and mortality of chronic kidney disease (CKD) and dialysis patients remain excessive, despite increasing knowledge on the role of a multitude of cardiovascular risk factors. As a hallmark of this, randomized controlled trials (RCTs) in these populations failed to demonstrate survival benefit of interventions successfully targeted on well-accepted risk factors in the general population or on dialysis efficiency [1]. Extensive research has confronted the nephrology community with a complex and elusive network of pathophysiological pathways. In contrast, our knowledge of basic elements, such as magnesium, remains surprisingly scarce. Nevertheless, emerging evidence, be it mostly circumstantial, from in vitro and in vivo studies, indicates that hypomagnesaemia, or rather a magnesium deficit per se might convey nephrotoxicity together with negative cardiovascular effects. Alternatively, higher serum magnesium levels might be beneficial. This review aims to further critically appraise and outline this premise. Of importance, the interventional trials targeting magnesium status in CKD and dialysis patients are almost non-existent. Confounding by unmeasured residual subclinical disease and multicollinearity could subsequently hamper the interpretability of the observational evidence. Hypomagnesaemia might merely be considered a ‘risk marker’ rather than ‘risk factor’. Future studies in the CKD and dialysis population therefore should meticulously address causality in order to allow the development of guidelines on optimal target values. Pham et al. [2] demonstrated in a cohort of predominantly Hispanic diabetes patients without known kidney disease (n = 550) that lower magnesium levels were associated with faster deterioration of kidney function, even after correction for age, lipid status and HbA1c. Hyperglycaemia is associated with enhanced renal magnesium wasting [3] and therefore could bias the observed relationship. Lower serum magnesium levels in diabetes patients, irrespective of glycaemic control, were independently associated with micro- and macroalbuminuria versus normoalbuminuria [4]. Also, unpublished observational data from our centre recently confirmed baseline serum magnesium to be a predictor of faster decline in kidney function in CKD patients (n = 1660) independent of baseline kidney function, age, gender, diabetes and hypertension. Hypomagnesaemia predicted a faster decline of kidney function in a retrospective analysis of renal transplant recipients with biopsy-proven cyclosporine toxicity (n = 60) [5]. Without adjustment for drug levels, higher drug exposure with more profound hypomagnesaemia might obscure the observed picture. However, in cyclosporine-treated rats on a low-sodium diet, magnesium supplementation ameliorated the hypomagnesaemia, protected against cyclosporine-induced kidney fibrosis and preserved the kidney function [6, 7]. This effect was associated with abolishment of the cyclosporine-induced increased cortical messenger RNA (mRNA) levels of fibrogenic and chemoattractant molecules like transforming growth factor beta, plasminogen activator inhibitor (PAI)-1 and monocyte chemoattractant protein 1 [6, 7]. Furthermore, magnesium inhibited the upregulation of renal cortical endothelin-1 mRNA expression [7]. Unfortunately, studies examining large registry databases, such as USRDS and NHANES, omitted serum magnesium levels as a covariate in their modelling of decline of kidney function. The following arguments, however, support adjusting for hypomagnesaemia in studies on development and progression of renal and cardiovascular disease (Figure 1, Table 1). The association of hypomagnesaemia and magnesium supplements with cardiovascular disease and renal injury The association of hypomagnesaemia and magnesium supplements with cardiovascular disease and renal injury The potential role of magnesium in both renal injury and cardiovascular disease. The grey lines demonstrate proven associations (irrespective of causality) with magnesium deficiency/hypomagnesaemia and the crosses represent points of action where magnesium has demonstrated beneficial effects in countering these risk factors or markers. The endothelium is the largest organ in the human body. It senses hormonal and mechanical stimuli and promotes the secretion of numerous compounds affecting vascular tone, inflammation, platelet activity and cellular proliferation. Endothelial dysfunction, which is already present in the early stages of renal failure, is an independent predictor of faster decline of kidney function in a prospective cohort analysis of 500 hypertensive subjects with normal or moderately impaired kidney function [8]. CKD is a state of reduced availability of nitric oxide (NO), the synthesis of which critical for endothelial function. Endothelial dysfunction and vascular inflammation are early and interrelated contributing factors in the pathophysiology of atherosclerosis [9] and predict cardiovascular events. They precede structural changes and, as such, seem plausible risk factors for enhanced progression of kidney disease. The relationship between magnesium and endothelium has been widely explored both in vitro and in vivo. In vitro, low magnesium concentrations reversibly inhibit endothelial proliferation and are associated with upregulation of PAI-1 and vascular cell adhesion molecule-1 (VCAM-1) [10] next to increased endothelial secretion of matrix metalloprotease (MMP)-2 and MMP-9 [11]. Exposure of endothelial cells to high magnesium concentration increases endothelial nitric oxide synthase (eNOS) expression [12] and suppresses the synthesis of the potent vasoconstrictor and mitogen endothelin-1, a promoter of progression of kidney disease in experimental animal models [13] and in proteinuric CKD patients [14]. Inbred mice with low intracellular magnesium have impaired endothelial function with decreased expression of eNOS and increased levels of pro-inflammatory markers [15]. In a placebo-controlled RCT in patients with coronary artery disease, magnesium improved endothelial function, especially in those with the lowest intracellular magnesium concentrations [16]. This finding was confirmed in diabetic patients aged >65 years, endorsing the causal relationship between magnesium and endothelial function [17]. It can be speculated that lower magnesium levels with their negative effect on endothelial function and endothelin-1 concentration might affect the rate of decline of kidney function or conversely that correction of hypomagnesaemia might be nephroprotective. In this hypothesis, it might seem counterintuitive that declining renal function is associated with progressive endothelial dysfunction, as it is likewise associated with increasing serum magnesium levels. This paradox is, however, explained by the presence of toxins accumulating in CKD with a much stronger effect on the endothelium such as asymmetric dimethyl arginine [1, 18] which predicts cardiovascular events in CKD/dialysis patients. Also, despite an increase in total serum magnesium with decreasing kidney function, the biologically active ionized fraction often is lower in dialysis patients than in the general population, partially through complexing of magnesium with phosphate and other anions [19]. In addition, serum magnesium only constitutes a small fraction of total body magnesium and poorly reflects the status of this largely intracellular cation in uraemic patients [20]. To conclude, a magnesium deficit, irrespective of serum magnesium levels, might exert negative effects on endothelial function and as such might contribute to renal injury and cardiovascular disease. Many observational studies have demonstrated an association between inflammation and endothelial dysfunction [21]. Inflammation is considered a risk factor of faster decline of kidney function [22, 23] and of increased cardiovascular morbidity and mortality in populations both with and without CKD. Also, markers of haemostasis such as serum von Willebrand factor (VWF) independently predict faster decline of kidney function in the general population [23]. Interestingly, levels of VWF increased with decreasing serum magnesium (P < 0.001) in the same cohort [24]. Magnesium has anti-inflammatory and anti-oxidant properties [25]. Conversely, a magnesium deficit in vitro increases the expression of the pro-inflammatory nuclear transcription factor nuclear factor kappa-B (NF-κB) in a dose-dependent way in both endothelial [11] and vascular smooth muscle cells (VSMCs) [26]. It also enhances the production of interleukin 1-beta and tumour necrosis factor-alpha in alveolar macrophages [27]. Furthermore, hypomagnesaemia is associated with increased oxidative stress, which is discernible already early in CKD [1]. Activity of endothelial cell NADPH oxidase is upregulated by low magnesium exposure [28]. In vivo, dietary induction of magnesium deficiency in rodents led to early leukocyte activation, release of pro-inflammatory cytokines and enhanced free radical production [29]. In a large cohort of non-diabetic adults aged <30 years (n = 4497), both low magnesium intake and a low serum magnesium were independently associated with increased high-sensitivity C-reactive protein (hs-CRP) [30]. In peritoneal dialysis patients, a lower serum magnesium level correlated both with higher hs-CRP levels and a poorer nutrition status [31] Insulin resistance, assessed by the homeostasis model assessment of insulin resistance (HOMA-R), is an independent risk factor for both deterioration of kidney function in CKD [32] and for mortality in dialysis patients [33]. In insulin-resistant rats, the glomerular insulin-induced phosphorylation of eNOS was selectively inhibited [34]. This might associate insulin resistance with glomerular endothelial dysfunction and damage in type 2 diabetes, metabolic syndrome and obesity; interestingly, all these conditions are conditions characterized by a high prevalence of hypomagnesaemia [3]. The relationship per se between hypomagnesaemia and insulin resistance has been studied extensively. It is a risk factor for the development of glucose metabolism disorders or diabetes in both the general population [3] and in transplant recipients [35]. However, reverse causation is a potential source of bias in these analyses. Hypomagnesaemia is believed to exert diabetogenic effects by modulation of insulin action and sensitivity [36], partially through modifications of tyrosine kinase on the insulin receptor level [3] or the induction of inflammation and oxidative stress. In healthy humans, a magnesium deficit through dietary restriction (<0.5 mmol/day) for 4 weeks increased insulin resistance [36]. Both dietary magnesium intake and serum magnesium levels inversely correlate with insulin resistance in a large cohort of young adults without diabetes [30]. Recently, magnesium supplementation increased insulin sensitivity in normo-magnesemic non-diabetic obese subjects in a placebo-controlled RCT [37], suggesting that hypomagnesaemia is not a prerequisite for generating favorable effects of magnesium supplementation. Sympathetic nerve overactivity is already detectable in early CKD. It is associated with the development of arterial hypertension, vascular remodeling and left ventricular hypertrophy amounting to an increased cardiovascular morbidity and mortality in dialysis patients [38]. However, CKD per se can trigger sympathetic overactivity [38] which can further aggravate proteinuria and contribute to progressive glomerulosclerosis, interstitial fibrosis and deterioration of kidney function [39]. Interestingly, a magnesium deficit induced sympathetic excitation in rats [40] while magnesium supplementation attenuated sympathetic nerve overactivity by inhibiting norepinephrin release through blockade of mainly N-type calcium channels [41]. Emerging evidence indicates that high aldosterone levels especially in combination with a high-salt diet are associated with increased progression of kidney disease and decline of function [42]. Aldosterone excess, which is common in obesity possibly through adipocyte secretion of mineralocorticoid-releasing factors, is associated with endothelial dysfunction, insulin resistance, inflammation, oxidative stress, hypertension, sympathetic nerve activation and increased extracellular matrix turnover and fibrosis [42, 43]. This is further supported by several studies pointing to nephro- and cardioprotective properties of aldosterone antagonists [43]. Interestingly, dietary induction of hypomagnesaemia in healthy humans increases aldosterone levels [36] while hyperaldosteronism leads to hypomagnesaemia and/or intracellular magnesium deficiency partially through renal magnesium wasting [44] and alterations of the Na+/Mg2+ exchanger [45]. Magnesium supplementation on the contrary attenuates the angiotensin-II-induced rise in serum aldosterone levels in healthy volunteers irrespective of their magnesium status [36]. Furthermore, magnesium supplementation counteracts the damaging cardiovascular and renal effects of aldosterone in both normo- and hypomagnesaemic mice [25, 44]. Aldosterone substantially increased renal and aortic collagen deposition, podocytal injury, the renal expression of VCAM-1, the phosphorylation of mitogen-activated protein kinases, oxidative stress and accordingly the activity of the redox-sensitive NF-κB in mice with hypomagnesaemia. These effects were attenuated by magnesium supplementation [25, 44]. Serum of hypomagnesaemic rats stimulated proliferation and collagen deposition in cardiac fibroblasts, which could be reduced by aldosterone blockade [46]. Spironolactone in this regard also increases intracellular magnesium levels in patients with heart failure [47], this in contrast with the effects of thiazides and loop diuretics. Also, in a crossover RCT in hypertensive patients, chlorthalidone unlike spironolactone increased both insulin resistance and sympathetic nerve activity [48], an effect possibly at least partially explained by a different magnesium status between both groups. To conclude, we can speculate that some of the negative and indirect pleiotropic effects of aldosterone are attributable to associated hypomagnesaemia. Conversely, hypomagnesaemia might exert part of its harmful effects through increased aldosterone production or through synergistic amplification of the effects of aldosterone. Vascular calcification is common in CKD and dialysis patients and is explained by both traditional and non-traditional risk factors, such as uraemic retention solutes and disturbed mineral metabolism. Emerging data have pinpointed the active role of transdifferentiation of VSMCs into osteoblast-like cells and of the unbalance between promoting and inhibitory proteins in this process. Based upon epidemiological and basic experimental findings, magnesium has been attributed anti-calcification properties (Table 2). In both uraemic and non-uraemic rats, a prolonged magnesium-deficient diet with subsequent hypomagnesaemia leads to increased aortic, cardiac and renal calcium concentration [49], widespread tissue calcifications [50], increased large artery media thickness and collagen content and higher pulse pressure and mortality [51]. In a cross-sectional analysis, increasing carotid intima–media thickness (IMT) was observed in women with decreasing serum magnesium levels after adjustment for potential confounders including diuretics [52]. Accordingly, in a Japanese cohort study, serum magnesium levels <2.2 mg/dL were independently associated with IMT and carotid plaques [53]. Also, in cohorts of both older renal transplant recipients and CKD Stage 5 patients, serum magnesium levels correlated with vascular stiffness, assessed by pulse wave velocity [54, 55] and IMT [55]. In peritoneal dialysis patients, the incremental evolution of peripheral arterial calcification >27 months was most pronounced in the cohort with the lowest magnesium levels, without any difference in serum calcium, phosphorus, calcium phosphorus product and parathyroid hormone (PTH) between progressors and non-progressors [56]. In haemodialysis patients, mitral valve calcification and increased carotid IMT were more prevalent in patients with lower (intracellular) magnesium levels [57, 58]; randomized allocation to magnesium citrate versus calcium acetate not only decreased PTH levels but interestingly also IMT [59]. Although not uniformly [60], various studies have demonstrated an independent relationship between low serum magnesium levels and increased cardiovascular and overall mortality, sudden cardiac death and stroke in, population-based cohorts or patients with advanced atherosclerosis [61, 62, 63]. One retrospective single-centre study in Japanese haemodialysis patients confirmed a higher but—remarkably—especially non-cardiovascular mortality in patients with lower serum magnesium levels [64]. Magnesium and vascular calcificationa aInterventional studies examining the role of magnesium in vascular calcification are depicted. BPG, β-glycerophosphate; mM, mmol/L; FMC, fetuin mineral complex; PXE, pseudoxanthoma elasticum; FGF, fibroblast growth factor; MGP, matrix Gla protein; TRPM7, transient receptor of potential melastatin 7; BMP-2, bone morphogenetic protein 2; NA, not applicable. Magnesium and vascular calcificationa aInterventional studies examining the role of magnesium in vascular calcification are depicted. BPG, β-glycerophosphate; mM, mmol/L; FMC, fetuin mineral complex; PXE, pseudoxanthoma elasticum; FGF, fibroblast growth factor; MGP, matrix Gla protein; TRPM7, transient receptor of potential melastatin 7; BMP-2, bone morphogenetic protein 2; NA, not applicable. A magnesium deficit associates with endothelial dysfunction, inflammation, insulin resistance, oxidative stress, hyperaldosteronism and hyperparathyroidism [19], all of which are associated with vascular calcification. Moreover, hypomagnesaemia is associated with traditional Framingham cardiovascular risk factors, such as diabetes, lipid disorders and hypertension [44]. In most studies, inverse correlations between (intracellular) magnesium levels and blood pressure are observed which are even more pronounced in patients with malignant hypertension or metabolic syndrome. Nevertheless, meta-analyses of interventional trials in the general population have demonstrated inconsistent results of magnesium supplementation on blood pressure control [65]. Low serum magnesium was a strong predictor of the 5-year increase in left ventricular mass in a large German cohort of patients >45 years after adjustment for many covariates including hypertension [66]. Also in animal models, magnesium attenuated aldosterone- or cyclosporine-induced harmful effects independent of blood pressure [6, 7, 25, 44]. Hypomagnesaemia induces an atherogenic lipid profile through activation of 3-hydroxy-3-methyl-glutarylCoA reductase (HMG-CoA reductase) next to a decrease of lecithin-cholesterol acetyltransferase and lipoprotein lipase activity [28], while magnesium beneficially alters lipid levels [67]. In vitro studies have pinpointed the anti-calcification effect of magnesium, which is a structural component of the calcium phosphate (apatite) mineral and a cofactor of alkaline phosphatase in especially cell culture models (Table 2). Higher magnesium exposure inhibited calcification in both bovine and murine VSMC [68, 69] while opposing their apoptosis and osteogenic transdifferentiation by downgrading transcription factors in a dose-dependent way. Interestingly, magnesium stimulated the expression of anti-calcification proteins such as osteoprotegerin [70], matrix Gla protein and osteopontin [68, 69]. Magnesium was able to inhibit progression of calcification and even reverse estabished lesions [69]. Magnesium exerted beneficial effects on phosphate-induced VSMC transformation in rats and on hydroxyl-apatite crystal formation in a more physiological rat model of isolated aortic segments [50]. Magnesium supplementation also protected against vascular calcification in a mouse model of pseudoxanthoma elasticum, a rare hereditary condition characterized by premature arterial calcification [71]. Possibly, the effect of magnesium, next to the effects on osteogenic transdifferentiation, relates to antagonizing calcium influx into the cell, and engagement of cellular cationic transporters especially transient receptor potential cation channels [69]. Additionally, in uraemic rat models, preliminary data also show that magnesium reduces the spontaneous precipitation of fetuin–mineral complex in a dose-dependent manner and delays the formation of stable calcium apatite, both suggesting a role for magnesium in the pathogenesis of uraemia-related calcification [72, 73]. Still many questions remain unanswered. For instance, the proven relationship between hypomagnesaemia and low 25-OH vitamin D levels could confound the acknowledged association between cholecalciferol, insulin resistance and vascular calcification. Also, the role of hypomagnesaemia as a surrogate marker for the prognostically ominous protein–malnutrition complex should be clarified. Although not all data are pointing in the same direction, most studies acknowledge an inverse relationship between serum magnesium and PTH levels in dialysis patients [74] with potential risk of adynamic bone disease which should be further explored. Finally, the potential threat of profound and potentially dangerous hypermagnesaemia should always be guarded, especially as renal failure progresses. Considering the alleged beneficial role of magnesium in CKD/dialysis patients, its role as a phosphate binder is being further outlined. In the CALMAG trial [75], an RCT featuring 255 haemodialysis patients, and the largest trial the effects of magnesium versus the phosphorus control and without increased effects. Although the observed small but rise in serum magnesium levels with this phosphorus binder might in a in with all the effects Of was to increase magnesium levels as in patients which might to the anti-calcification potential of the Considering the evidence in the the further of the role of serum or more intracellular magnesium in CKD or dialysis patients further The data to an for further research and as observational data This might to a and the of magnesium deficiency as a risk marker or even a risk factor for cardiovascular disease and decline of kidney function in CKD with a potential for of