Calcium-based phosphate binders; down, but not out Free

In 1960, Dr Bricker and colleagues described how serum phosphate could be maintained within the normal range despite nephron destruction, by an increase in phosphaturic hormones acting on remaining, intact nephrons [1]. However, this adaptive response was not required if phosphate intake was limited simultaneously with the reduction in residual renal function [2]. Subsequently, dietary phosphate restriction and phosphate-binding drugs have been used almost universally to avoid the potential complications of hyperphosphataemia in chronic kidney disease (CKD). Calcium-based phosphate binders (CBPBs) seemed particularly suitable, because they avoided the toxicity of aluminium, directly suppressed parathyroid hormone (PTH) and were cheap. However, concerns for positive calcium balance have led many clinicians to avoid calcium, or to use CBPBs in combination with newer, more expensive non-calcium-based drugs.

Despite concerns for the consequences of positive calcium balance, surprisingly few studies have compared any phosphate binder with placebo, compared outcomes of treating with differing phosphate targets, or have had the power and duration to provide important patients-level outcomes. Consequently, a number of meta-analyses have interrogated the available data [3–7], the latest of which, published in this issue of Nephrology Dialysis Transplantation, compares sevelamer and lanthanum with other predominantly calcium-based binders [8].

To summarize the finding briefly, compared with CBPBs, mortality was 38% lower for sevelamer [relative risk (RR) 0.35, 1.08] and non-significantly lower for lanthanum. Hospitalization rates were lower for sevelamer and hypercalcaemic events occurred 3- to 4-fold less often with sevelamer or lanthanum than CBPBs. Lower end-of-study calcium values were associated with reduced mortality, but no such association was detected for phosphate values. Low-density lipoprotein-cholesterol values and coronary artery calcification (CAC) scores were lower for patients treated with sevelamer than CBPBs, and patients treated with CBPBs had greater increases in CAC scores, but these endpoints were not associated with reduced mortality. No differences were detected in three studies comparing sevelamer with iron-based binders for serum calcium, phosphate, PTH or adverse events.

The author’s concerns are similar to those expressed in recent comparable analyses; in particular that study heterogeneity and risk of bias, periods of follow up ranging from 2 weeks to 3 years, incomplete outcome data and selective reporting might erode confidence in significant differences. While it is suggested that recent reviews failed to adequately address important outcomes such as cardiac events, bone fractures, hypercalcaemia and all-cause mortality, this ‘inadequacy’ is due to a paucity of data. Those outcomes and others, including quality of life, levels of serum fetuin-A and fibroblast growth factor 23 (FGF23), have recently been sought [5]. Similarly, some criticisms of studies omitted in earlier meta-analyses in fact relate to predetermined inclusion/exclusion criteria; for example, studies of <8 weeks duration, patients with CKD stage 2, post hoc analyses and studies in which authors did not explicitly report numbers of deaths were excluded a priori from a recent systematic review [5].

Looking at recent meta-analyses, one might reasonably question the basis for phosphate measurement, treatment to target ranges, or whether these studies have assisted clinicians in making decisions on whether CBPBs or non-calcium-based drugs might be preferable for individual patients. This has not been assisted by headlines such as ‘The demise of calcium-based phosphate binders’ [9] that have editorialized some contributions. However, recent meta-analyses and the studies upon which they are based do provide some points of consensus:

• Mortality associated with abnormalities of mineral metabolism increases with each stage of CKD and abnormal values of serum phosphate and calcium are predated by initially homeostatic, but progressively maladaptive changes in levels of FGF23 and PTH.

• For patients not on dialysis, it is uncertain whether any phosphate binder improves patient-level outcomes. For patients on dialysis, the value of phosphate-binding drugs is more established, but targets for serum phosphate remain controversial.

• CBPBs are not superior to non-calcium-based binders for any patient-level outcome, although their lower cost remains an important consideration.

• Treatment with CBPBs contributes to positive calcium balance and hypercalcaemia. If non-CBPBs have an advantage, it is probably by avoiding this consequence. However, some non-CBPBs may have additional beneficial pleiotropic effects on lipids, inflammation, FGF23 and fetuin-A, and in the case of ferric citrate, reduced requirements for iron supplementation and erythropoiesis-stimulating agents (ESAs) [10].

THE BIG QUESTION: IS THERE A CASE FOR PHOSPHATE LOWERING IN CKD?

For patients with CKD, extensive observational data supports the association of elevated serum phosphate to the prevalence and severity of CAC [11] and to mortality [12–14]. However, the value of interventions to reduce serum phosphate is less clear. In the pre-dialysis population, the use of phosphate binders is particularly controversial, because when the effect of phosphate binders was studied in moderate CKD [15], the risk of developing vascular calcification was greater for patients allocated to lanthanum, sevelamer and calcium-based binders than to placebo.

FGF23 reduction is another potential benefit of treatment with non-CBPBs. Elevated levels of FGF23 appear to influence cardiac fibrosis, hypertrophy and arrhythmogenic potential through calcium dysregulation [16, 17], and higher levels have been associated with greater risks for death and progression to dialysis in patients with CKD stages 2–4 [18]. FGF23 reductions have been reported in CKD stages 3–4 after treatment with sevelamer hydrochloride over 6 weeks [19] and ferric citrate over 12 weeks [20], and in CKD stages 4–5 after treatment with lanthanum over 4 months [21]. These interventions with non-CBPBs have potential to delay initially adaptive FGF23 responses from becoming detrimental. The current long-term IMPROVE study of lanthanum versus placebo in CKD stages 3–4 [22], may assist in determining the success of early intervention strategies.

For patients on dialysis, a number of non-interventional, observational studies have reported that any phosphate binder therapy, apart from aluminium, is associated with decreased mortality; by up to 42% over the first year of dialysis [23] and in COSMOS (the Current management Of Secondary hyperparathyroidism: a Multicenter Observational Study), by 29% after multivariate adjustment, over 3 years [24]. However, in an analysis of patients enrolled in the Dialysis Outcomes and Practice Patterns Study (DOPPS), phosphate binder prescription was strongly associated with nutritional status, and adjustment for nutritional indicators attenuated the association of phosphate binders to improved survival [25].

PHOSPHATE DIRECTLY ACTIVATES PATHWAYS INFLUENCING CKD–MBD DEVELOPMENT

Compelling data support the concept that elevated phosphate levels influence the development and progression of CKD–MBD (Mineral and Bone Disorder), and recent studies have suggested additional novel mechanisms. In vascular smooth muscle cells (VSMCs), elevated serum phosphate levels promote vascular calcification by increasing the activity of type III sodium-dependent phosphate co-transporters (PiT-1 and PiT-2), which mediate the transdifferentiation of VSMCs to cells with osteoblast characteristics [26–28]. In fact, recent studies suggest that PiT-1 sodium-dependent phosphate co-transporters in VSMCs and the parathyroids may be ‘transceptors’, having the capacity to both transport and sense extracellular phosphate [29, 30]. Another recent finding based on structural analyses is that the calcium-sensing receptor (CaSR) may be a phosphate sensor [31]. Extracellular phosphate is reported to reinforce the CaSR’s inactive conformation to facilitate the production of PTH.

CONCLUSIONS

There is a strong epidemiological and physiological basis for concern over elevated levels of serum phosphate in patients with CKD. Despite a lack of randomized controlled trials to assess the relative benefit of targeting differing serum phosphate values, most data supports current KDIGO (Kidney Disease: Improving Global Outcomes) guidelines that suggest we should aim for phosphate values within or towards the normal range [32].

Given the marginal theoretical advantages of one non-CBPB over another, it is unlikely that head to head trials comparing sevelamer with lanthanum or with other non-calcium-based binders will demonstrate differences in mortality, although for assessing adverse events, pill burden, compliance and effects on anaemia, ESA use and economic costs, such comparisons may be worthwhile. Neither will additional studies of this sort illuminate our understanding of calciphylaxis or fracture risk. We are more likely to progress those areas by supporting the development of calciphylaxis registries, and by improving the biochemical assessment of bone turnover and imaging techniques, than investing in studies comparing phosphate binders. But we should not give up the possibility that trials comparing regimens that reduce phosphate ‘balance’ may provide valuable patient-level outcomes, particularly studies that assess non-CBPBs combined with drugs such as nicotinamide that reduce active phosphate absorption by gastrointestinal sodium-dependent phosphate transporters type IIb [33]. Such combinations could reduce phosphate exposure and adverse effects, by enabling lower doses of each component of the regimen.

Despite CBPBs increasing the risk for outcomes such as hypercalcaemia and CAC, which most clinicians wish to avoid, the current meta-analysis concludes that we ‘cannot be definitive about whether sevelamer (or lanthanum) reduces, has no impact, or increases the risk of death’. While this may be a statistically defensible statement, the KDIGO 2016 Clinical Practice Guideline Update on Diagnosis, Evaluation, Prevention and Treatment of CKD–MBD that is currently undergoing public review, suggests that ‘In adult patients with CKD Stages 3a-5D receiving phosphate-lowering treatment, the dose of calcium-based phosphate binders be restricted (Grade 2B) and in adult patients with CKD Stages 3a-5D, hypercalcemia should be avoided (Grade 2C) (http://www.kdigo.org/clinical_practice_guidelines/CKD-MBD%20Update/KDIGO%20CKD-MBD%20Update_Public%20Review_Final.pdf, 22 October 2016, date accessed). These suggestions are considered, and consistent with the available physiological and study data. Given that there seems little to recommend CBBPs apart from cost, a reasonable conclusion is that calcium is down, but not out.

CONFLICT OF INTEREST STATEMENT

Comments in this editorial have not been presented in whole or in part elsewhere. G.J.E. has been a member of the advisory boards of Sanofi and Shire Australia, and has received speakers fees and travel grants from both companies.

(See related article by Habbous et al. The efficacy and safety of sevelamer and lanthanum versus calcium-containing and iron-based binders in treating hyperphosphatemia in patients with chronic kidney disease: a systematic review and meta-analysis. Nephrol Dial Transplant 2017; 32: 111–125)

REFERENCES