Getting Out of the Phosphate Bind: Trials to Guide Treatment Targets

• Biophysical Phenomena
• Phosphates

Treatment of serum phosphate has been a mainstay in the care of patients with ESRD on hemodialysis (HD) practically since its inception. Initially, treatment was justified on the basis of phosphate balance, with the recognition that phosphate removal by thrice-weekly dialysis was inadequately matched to dietary intake (1). Growing evidence of epidemiologic associations between circulating phosphate and mortality and adverse vascular effects of phosphate in experimental models strengthened this paradigm (2,3). With this rationale, intensive phosphate management with binders was widely adopted in practice and bolstered by clinical guidelines that advocated near-normalization of circulating phosphate on the basis of low-quality evidence (4,5). Unfortunately, mortality among patients on HD has remained stubbornly high (www.usrds.org), causing many to question the efficacy and possible opportunity costs of an intense focus on phosphate management in the HD unit. According to Medicare Part D data, phosphate binders are among the most widely prescribed dialysis-related drugs in the United States (6), in use by >80% of patients (the Dialysis Outcomes and Practice Patterns Study [DOPPS] Practice Monitor; www.dopps.org). Despite this ubiquity, there are essentially no adequately powered clinical outcome studies determining the effectiveness of phosphate binders or the benefits of phosphate reduction more globally.

Where conducted, most prior outcome trials in phosphate management have evaluated the comparative effectiveness of different binder classes, leaving the question of overall effectiveness unanswered (7). This void largely reflects the challenge of conducting placebo-controlled trials after a standard of care has been broadly accepted. For many analogous risk factors, such as BP, trials have led to increasingly intensive treatment targets over time (8). In contrast, liberalizing established treatment targets for phosphate will require consensus in the field that the established standard is poorly justified and that randomization to targets outside of guideline recommendations is ethical. Whether patients and providers will accept and achieve these targets is unknown and a potential barrier to conducting needed outcome trials.

In this issue of the Clinical Journal of the American Society of Nephrology, Wald et al. (9) provide such evidence in a pilot randomized, controlled trial evaluating the feasibility of intensive versus liberalized phosphate goals and help usher in a new era of evidence-based phosphate management. In the Two Phosphate Targets in ESRD (TARGET) Trial, they studied 104 adult patients with prevalent ESRD from five centers across Canada with a median dialysis vintage of 3 years. All patients were receiving dialysis up to four times weekly and had a serum phosphorus of 4.03–7.75 mg/dl (1.3–2.5 mmol/L phosphate), despite use of calcium carbonate with meals. Participants were randomized to intensive (goal serum phosphorus: 2.33–4.66 mg/dl or 0.75–1.50 mmol/L) or liberalized (goal serum phosphorus: 6.20–7.75 mg/dl or 2.0–2.5 mmol/L) targets for a 26-week period, with phosphate management achieved via calcium carbonate withdrawal or titration. The small number of participants also using noncalcium-based binders at baseline could continue use, but they were not titrated. Dialysis time was not changed during the study.

At completion, serum phosphorus decreased by a mean of 1.24 mg/dl (0.40 mmol/L phosphate) in the intensive compared with the liberalized group. The study was powered to detect this difference in serum phosphate between groups but not clinical outcomes. There were no meaningful changes in intact parathyroid hormone, dialysate calcium prescriptions with HD, or the dose of activated vitamin D analogs between arms, and there were no differences in patient-reported quality of life or clinical outcomes.

Although it may seem obvious that more aggressive titration of phosphate binders would effectively lower phosphate, there are important lessons in the details. For instance, although there was good separation of serum phosphate across groups on average, many individual patients did not obtain the treatment target. In the liberalized group, only 27.5% of participants achieved a phosphorus >6.20 mg/dl (>2.0 mmol/L phosphate), with the achieved mean level in the liberal arm only minimally higher than current guideline ranges. If future studies aim to test the noninferiority of more liberal phosphate management, it will be important for an adequate number of individuals to achieve the liberal target levels. If not, the safety of this more liberal approach could remain in question. A phosphate binder washout period before randomization would facilitate selection of participants who achieve the liberal target range off binders and help ensure that randomized liberal targets are met.

Surprisingly, approximately one half of patients in the intensive group of the TARGET Trial achieved their phosphate targets with calcium carbonate alone, and most participants in the liberal phosphate target group achieved their target or lower without use of binders at all (9). This may not reflect the dialysis population more broadly, where use of two high-dose binders is often necessary (10). The challenges in achieving goal phosphate ranges reminds us that phosphate itself is a complex target. In addition to binders, phosphate is influenced by diet, dialytic clearance, intracellular shifts, circadian variation, and additional treatments, such as vitamin D and calcimimetics. Studies in humans and animals show plateauing and even rising of serum phosphate concentrations during the later phase of dialysis (11,12). Phosphorus magnetic resonance imaging in pigs suggested that this rise is due to inorganic phosphate generation or release from an intracellular source, where the vast majority of phosphate is sequestered (11). Studies such as these challenge a simplistic view of serum phosphate as a barometer of external phosphate balance and help emphasize why outcome trials are desperately needed. Although the risk associated with higher phosphate is remarkably consistent in the observational literature, these studies predominantly evaluate phosphate levels, not treatments (3). Trials that randomize the approach to treat phosphate intensively or liberally, as piloted here, could plausibly deviate from observational results if the risk associated with higher phosphate is not causal and instead, is a marker of difficult to measure health variables, such as intracellular or bone phosphate mobilization.

Randomizing the phosphate approach by focusing on the target level assumes that the primary benefit of phosphate binders is via overt phosphate reduction. However, phosphate binders may affect more than circulating phosphate concentrations, including phosphate regulatory hormones and other risk factors (13,14). Recent meta-analyses show poor correlation between achieved phosphate levels and outcomes in small trials to date (7). Alternative trial designs that randomize participants to different drug doses, such as placebo or low- or high-potency phosphate binding, could more directly evaluate the effects of treatment with binders. This paradigm would align with changes in management of other risk factors, such as lipids, in which newer guidelines advocate management of statin intensity according to drug and dose as opposed to the prior focus on low-density lipoprotein cholesterol targets (15).

Furthermore, although evidence in the field is generally limited, some meta-analyses support better outcomes with noncalcium- versus calcium-based phosphate binders (13,16). In this trial, all patients were treated with calcium carbonate as a binder. Although calcium-based binders are more common in Canada, where the study was conducted, noncalcium-based binders predominate in the United States (the DOPPS Practice Monitor; www.dopps.org). Binding potency, tolerability, and cost of calcium- and noncalcium-based binders differ (10,13). Furthermore, alternative payment models in the United States may affect adherence to binders and limit the achievable difference (6). Pilot studies in the United States and other practice environments may help establish the broader generalizability of this randomized study.

Despite the aforementioned limitations, the authors have taken a critical step toward evidence-based phosphate management in ESRD. They have conclusively shown the feasibility of achieving distinct phosphate targets rapidly and sustainably, with a difference in serum phosphate across groups that is anticipated to be clinically meaningful. They use a pragmatic design with a low-touch intervention that should scale to a much larger outcomes trial. Most importantly, they remind us that, although the rationale for a plausible benefit of phosphate binding in ESRD is sound, definite outcomes studies are critically needed. Effectively testing the dogma surrounding phosphate management in HD requires a willingness of the nephrology community to consider liberalizing phosphate management, thereby enabling large controlled studies and avoiding the treatment crossover and dropout that has hampered other trials in mineral metabolism (17,18). This study shows that distinct phosphate targets are feasible and provides hope for these definitive studies. The nephrology community should strongly support efforts to develop evidence for phosphate management by conducting appropriately powered outcomes trials.

• Copyright © 2017 by the American Society of Nephrology