Hyperphosphataemia is recognised as an independent CV risk factor. Abnormal P metabolism occurs early in CKD and there is a general consensus that it is one of the most important factors contributing in the onset of CV calcification, along with changes in intra- and extracellular Ca content. Both have a strong influence on the vascular smooth muscle cell (VSMC) function.27,28 Since the experiments by Jono et al.27 and Giachelli et al.,29 nephrologists have recognised the need to avoid P overload, not only as a promoter of secondary hyperparathyroidism, but also because of its direct effects on CV health,30–32 including its potent proinflammatory and oxidative effects33,34 that may affect even patients with mild CKD and possibly the general population.35

Goodman et al.’s initial publication in adolescents on dialysis, showed that they already had vascular calcifications and that the amount of Ca they ingested was double than those without calcifications. This information started a broad debate about the different binders (calcium vs. non-calcium) that has not been fully resolved yet.36–38 Several randomised studies conducted in adults on dialysis have demonstrated that CV calcification progression was truly modifiable by choosing non-calcium P binders.7,39–42 However these results are not consistent.43 In some clinical trials, negative results have been attributed to a patient population that had a higher number of CV risk factors;44,45 other studies faced obstacles such as a small sample size or the use of high Ca concentrations in the dialysate bath in many patients.46 Therefore, despite the fact that the “Dialysis Clinical Outcomes Revisited” (DCOR) study found that using sevelamer HCl in dialysis patients did not significantly improve the mortality rates,47,48 the 2009 KDIGO guidelines proposed a restriction of Ca-based P-binder dose in the presence of arterial calcification (guideline 4.1.5; 2C), at least until more conclusive studies are conducted.30 This represents a step forward in comparison to the previous American guidelines49 (K-DOQI 2003) where, curiously, the Ca-based P binders were limited only to cases with severe vascular calcification (a situation that may be too advanced to modify the harmful consequences). Although the DCOR study did not strictly demonstrate the superiority of sevelamer versus calcium-based binders, it cannot be said that Ca-based P binders are safe.32,50 With need to take into consideration also economic aspects,51 at least these studies have made the nephrology community aware that indiscriminately using calcium-based binders may be inappropriate and that it might be safer to limit Ca intake to less than 1g/day in CKD patients.52 A recent metabolic study demonstrated that even in stage 3b-4 CKD patients (n=8; mean GFR 36ml/min/1.73m2; mean P=3.8mg/dl) administering 1.5g of calcium carbonate converted a neutral Ca balance into a largely positive balance,53 although it is unknown whether this altered balance is temporal, by inducing adaptation phenomena, or whether this excess of Ca ends up being deposited in extraosseous tissues.

More recent studies in dialysis patients have shown that the progression of vascular calcification was attenuated not only by sevelamer but also bylanthanum and, in two small pilot studies, by the use of P binders containing magnesium.54–59 A recent meta-analysis has also reinforced the idea that it may be possible to attenuate the progression of vascular calcification with non-calcium P binders.60 Moreover, in another randomised, open-label, parallel groups study including 466 Italian patients who were starting haemodialysis, sevelamer improved survival as compared with calcium-based P binders,61 although it could not be established a direct relationship between the Ca load and worse outcomes.61 Furthermore, in another recent meta-analysis, non-calcium P binders were associated with a criticised reduction of overall mortality risk (22%) in CKD patients (mostly dialysis patients treated with sevelamer),4 in contrast to the negative results from other previous meta-analyses.37,62 Nonetheless, in some of these studies high doses of binders were administered to achieve protocol objectives; therefore, these results should be extrapolated with caution to those situations in which moderate doses of Ca-based binders are used or those cases in which both types of binders (with and without Ca) are being administered.63,64 Another study found that normal individuals and patients with stage 3b-4 CKD had a slightly negative or neutral Ca balance while eating a 800mg/day Ca diet, however with a 2000mg/day diet, the normal individuals had a slightly positive balance and the CKD patients had a clearly positive balance, at least during the 9 days of the study.65

Lastly, a randomised, multi-centre, open-label pilot study conducted in 212 outpatients with stage 3–4 CKD recruited over a maximum of 36 months demonstrated that treatment with sevelamer to maintain plasma P within the normal range, was associated with a significantly lower incidence of de novo CAC among patients with no baseline CV disease (12.8% vs. 81.8% for sevelamer and Ca carbonate, respectively), as well as slower CAC progression among the patients with evidence of CAC at the start of the study.66 A significant regression of CAC was also detected in 24 patients treated with sevelamer, and only in 2 patients treated with Ca. The overall mortality and the final composite endpoint of death and dialysis inception were lower in the patients assigned to sevelamer.66 This study did not include a placebo arm and included patients with moderate hyperphosphataemia (4.84±1.3mg/dl). Conversely, in another, smaller randomised study of 148 patients with moderate CKD (GFR=20–45ml/min/1.73m2 with a mean P of 4.2mg/dl) comparison of Ca phosphate binders, sevelamer, and lanthanum vs. placebo showed a completely unexpected increase in vascular calcification in all groups, even though in the post hoc analysis, the degree of progression was higher in the Ca arm.67 In another study,68 rosuvastatin and sevelamer did not delay the progression of vascular calcification in CKD patients not yet in dialysis. Therefore, despite the demonstrated potential benefit for survival, at least in some CKD patients, additional studies are needed to define the effects of P binders in CKD patients before starting dialysis. In fact, this is one of the areas of nephrology in which we have less evidence and limited alternatives. A large controversy has been generated on whether P binders should be prescribed in stage 3–4 CKD,69–71 this controversy illustrate the extreme necessity to conduct prospective clinical studies measuring hard events.69–71 According to the prescribing information, P binders without Ca should be restricted to CKD patients not in dialysis if serum P is greater than 1.78mmol/l (5.5mg/dl). Nevertheless, it should be noted that in the recent controversies on the KDIGO guidelines,63 the concern about Ca overload as a risk factor for progression of vascular calcification in CKD was clearly emphasised.37,66,67,72,73

In addition to the experimental data showing a neutral or protective effect of calcimimetics on uraemic atherosclerosis or vascular calcification,24,74,75 a randomised clinical trial recently demonstrated that cinacalcet, along with low-doses of vitamin D analogues, can attenuate the progression of vascular calcification in dialysis patients versus the standard treatment (different doses of vitamin D analogues or binders).25 Although the study did not clearly demonstrate a significant benefit (p=0.07), there was a clear tendency towards decreased progression of CAC, thoracic aorta calcification, and cardiac valve calcification in the group treated with calcimimetics. This effect was especially significant in those patients adhering to the initially designed protocol maintaining a low dose of vitamin D analogues.26 Similar results were described in an observational study conducted in Japan.76 Furthermore, another retrospective study including dialysis patients on intravenous vitamin D therapy (a surrogate marker for secondary hyperparathyroidism), the prescription of calcimimetics was associated with a significant improvement in survival.77 However, the EVOLVE study,78 comparing calcimimetics vs. standard therapy in the largest study conducted in haemodialysis patients (3883), showed that cinacalcet did not significantly reduce the risk of death or CV events in dialysis patients with moderate to severe secondary hyperparathyroidism after an unadjusted intention-to-treat statistical analysis. Similarly, in a recent meta-analysis, based essentially on the above study, calcimimetics did not seem to improve CV or overall mortality.79 As a result, as previously observed with sevelamer in the DCOR study,47,48 it was not possible to definitively establish a direct relation between therapeutic measures that potentially attenuate vascular calcification progression in dialysis patients and benefits for survival. However, it is important to note that, in addition to other nominally significant beneficial effects associated with sevelamer and cinacalcet,47,48,78 it was observed in both studies that age had a highly significant interaction on the treatment effect. Both drugs significantly reduced mortality in a predefined subgroup of patients over 65 years of age; this result is likely due to the higher statistical power inherent to a higher number of CV events and mortality in this age group.78,80 A similar interaction with age was observed with lanthanum carbonate.81 In addition, in the general population a significant association between vascular calcification and kidney function has also been recently described in the elderly, but not in younger individuals.82 It is important to note that, in the case of the EVOLVE study, cinacalcet did significantly reduce the risk of death or major CV events in dialysis patients in a second predefined intention-to-treat analysis when adjusted for age or other factors, as well as other complex sensitivity analyses, despite the excessive number of drop-ins and drop-outs.78,83 Various beneficial effects of cinacalcet have also been described in post hoc studies,83–88 including decreased mortality in non-atherosclerotic events (including sudden death and heart failure) in patients treated with cinacalcet.86 We therefore believe that the EVOLVE study should not be considered a negative study, but rather an inconclusive, non-definitive study, since the absence of evidence cannot in any way be considered evidence of absence.50,83

Lastly, it is important to emphasise that calcimimetics have also been successfully used to treat some cases of calciphylaxis.87,89 Calciphylaxis episodes occurred significantly less often in the group of patients treated with cinacalcet in the EVOLVE study and in a post hoc analysis.78,87

Low levels of calcidiol (25-OH vitamin D) have been directly associated with the presence and progression of vascular calcification and represent a new CV risk marker on their own.90,91 Although it may only be a mere bystander maintaining “normal” levels seems to be desirable. Spanish guidelines92 recommended to maintain normal levels of calcidiol to reduce vascular calcification progression, maintain a normal bone turnover,6,91 and provide the other pleiotropic effects described for vitamin D, including vascular regeneration, anti-inflammatory effects, and anti-renin activity, among others.93–96 However, there are no prospective, randomised clinical trials that have assessed the impact of native vitamin D or vitamin D receptor (VDR) activators such as calcitriol, alfacalcidol, paricalcitol, or others on human vascular calcification. Experimental studies have demonstrated differential effects between calcitriol and other VDR activators on extraosseous calcification. Calcitriol is a classic, direct, dose-dependent inducer of experimental vascular calcification, especially in the presence of high P exposure or as the result of systemic vitamin D-induced Ca and P accumulation, more than a local effect on the artery wall.24,74,97 Furthermore, the lowest doses of both calcitriol and paricalcitol seem to protect against vascular calcification, likely through klotho restoration and osteopontin expression.98–100 Therefore, the presence of a bimodal effect of the VDR activators regarding vascular calcification regulation can be postulated. In general, experimental data supporting lower toxicity with some VDR activators versus calcitriol are not consistent between studies, but they seem to support the assertion that there is reduced calcification induction with other selective VDR activators such as paricalcitol.24,74,95,101 For example, paricalcitol, in contrast to calcitriol, decreases Wnt/β-catenin pathway activation, the most important signalling pathway in transdifferentiating VSMC into osteoblasts.102 Paricalcitol may also have an effect on earlier stages of vascular disease; it is unknown whether this is true of other VDR activators.103 Furthermore, several retrospective studies have described a consistent and solid benefit on survival for haemodialysis patients with selective VDR activators104,105 and, although it has been questioned,106 the benefit seems to be more pronounced in the low-dose range and among patients who received selective VDR activators.18,104 Lastly, a recent meta-analysis including 14 observational studies (194,932 patients) showed that VDR activator therapies are associated with a lower mortality in CKD patients,107 although there is no consistency between the different meta-analyses.108 To date there is no published prospective clinical trial assessing the effect of VDR activators on survival so the previous potential beneficial results could be confirmed, although by no means this beneficial effect should be rejected either.109

Vitamin K is necessary as a cofactor in the process of converting inactive decarboxylated extracellular matrix proteins into active carboxylated proteins. Osteocalcin and Matrix Gla protein (MGP) require the presence of vitamin K for activation and warfarin, as a vitamin K antagonist, inhibits coagulation, but long-term use can promote vascular calcification and overregulation of decarboxylated MGP.110,111 The association between CAC and vitamin K antagonist therapy was already known in patients with low-risk atrial fibrillation112 and, recently, Górriz et al. confirmed the independent association between the use of oral anticoagulants and vascular calcification, even in CKD patients not on dialysis.113 Experimental work shows that vitamin K is able to revert warfarin-induced medial calcinosis of elastin114 and, since vitamin K deficiency is common in dialysis, it is not surprising to see that currently there are several prospective clinical trials evaluating the effect of vitamin K supplementation on CAC progression in CKD and haemodialysis patients.115 It is possible that the new oral anticoagulants, now available for patients with atrial fibrillation or acute coronary syndrome, may become a therapeutic alternative.116,117

Bisphosphonates have also been successfully used “off label” to treat calciphylaxis.118 In addition to the experimental data showing that treatment with pamidronate or etidronate prevents vascular calcification,119 oral or parenteral etidronate can delay CAC progression and aortic valve calcification, although not all new-generation of bisphosphonates have been shown to reduce calcifications.120–123 In this line, it is worth to mention that the vessel wall possesses a “natural form of bisphosphonates”, pyrophosphates, which antagonise alkaline phosphatase and are one of the most effective anticalcifying factors of the vascular wall.

A randomised clinical trial including 108 hypercholesterolaemic patients revealed that combination therapy with atorvastatin plus etidronate for 12 months significantly reduced atheroma plaques in thoracic and abdominal aorta.124 Since the vascular effects of bisphosphonates cannot be separated from adequate bone formation, the administration to CKD patients may promote the development or aggravation of adynamic bone disease.30,125,126 As a result, a bone biopsy is recommended before using bisphosphonates in patients with a GFR<30ml/min/1.73m2 unless a high-turnover bone disease is undeniably present,30 or in the context of a potentially fatal disease such as calciphylaxis.118 A similar strategy should likely be applied to new therapies such as denosumab and romosozumab, although the half-life of these drugs in bone is certainly lower.127

Sodium thiosulfate has been introduced into the therapeutic arsenal against calciphylaxis.128 It may also attenuate the CAC progression rate versus the non-treatment group, but with a significant decrease in hip bone mineral density.129,130 Sodium thiosulfate and other binding agents have been demonstrated to be potentially useful in reversing vessel medial calcification,131 however the mechanism by which sodium thiosulfate reduces calcification is not fully understood.129,132

Given the importance of this topic, new drugs are being developed that could act as vascular calcification inhibitors such as SNF472,133 an intravenous formulation of myoinositol hexaphosphate (phytate) that prevents hydroxyapatite crystals from forming and growing.134,135 It also acts as a calcification antagonist that could be effective as a therapy for treating CV calcification in CKD patients and in calciphylaxis.133 SNF472 acts through a physicochemical mechanism, binding to the forming or growing crystal.136 Its high efficacy in animal models and short half-life give it a suitable safety and efficacy profile in CKD, but this will have to be confirmed in long-term clinical studies. At this time it is in phase 1b/2 development.136

There are no studies investigating the effects of parathyroidectomy on the progression or regression of vascular calcification that meet the pre-established inclusion criteria for the 2009 KDIGO guidelines revision. Similarly, to date there are no new data available beyond the classic indication for parathyroidectomy in the form of calciphylaxis associated with severe secondary hyperparathyroidism. In kidney transplantation, few studies have been able to demonstrate stabilisation or attenuation, but do not completely stop, the rate of progression of vascular calcification, despite the significant improvement in kidney function and mineral metabolism parameters.137–140 However, many other CV risk factors, either prior to or within the context of transplantation, may play an additional role in this specific population.