2.1 The KDIGO 2017 reframing

KDIGO 2003 / K/DOQI 2003 specified phosphate targets of 3.5–5.5 mg/dL in dialysis patients with calcium-phosphate product < 55 mg²/dL² [1]. These numerical targets drove a generation of dialysis quality programs and pay-for-performance metrics. The KDIGO 2017 CKD-MBD Update (Recommendation 4.1.1) deliberately moved away from strict numerical targets and toward a directional framework: “in patients with CKD G3a–G5D, we suggest lowering elevated serum phosphate levels toward the normal range” (level 2C, weak recommendation, low-quality evidence) [1].

The change was driven by an honest reading of the trial evidence — there was none demonstrating that hitting any specific phosphate number improved survival, cardiovascular events, or fracture. The framework retained the clinical priority (treat hyperphosphatemia) but abandoned the false precision of a hard cut-point.

Operationally, most dialysis units still treat 5.5 mg/dL as the actionable threshold, both because it remains the regulatory anchor (ESRD QIP performance scoring) and because the observational mortality signal does steepen above approximately 5.5 mg/dL [2,3].

2.2 ESRD QIP, KCQA, MIPS — what gets measured and why

ESRD QIP (CMS): Hypercalcemia (3-month rolling mean total serum calcium ≥ 10.2 mg/dL) is the directly-measured mineral metabolism quality metric in current ESRD QIP [1]. Hyperphosphatemia is not currently a direct performance measure but appears in clinical performance reporting through transparency mechanisms. The pragmatic effect: facility medical directors monitor phosphate-control percentages because they shape bundled-payment performance and Five-Star ratings even when the metric is not explicitly punitive.

KCQA (Kidney Care Quality Alliance): Has historically endorsed phosphate-related dialysis quality measures; the current KCQA core set continues to include mineral metabolism markers.

MIPS (Merit-Based Incentive Payment System): Nephrology MIPS measures touch CKD-MBD primarily through CKD progression and process measures (e.g., serum phosphorus monitored at appropriate intervals in CKD G4–G5). Outcome-based phosphate measures are not currently part of the MIPS nephrology specialty set.

2.3 Why the upper bound shifted

Three pieces of evidence converged. First, the observational mortality signal at the high end of phosphate distribution proved robust across multiple cohorts (Block 2004 [2], DOPPS [4], COSMOS [5]). Second, attempts to demonstrate that lowering phosphate with binders improves hard outcomes produced consistently null or marginal results (DCOR [6], INDEPENDENT [7], multiple meta-analyses). Third, the lower-bound concern — that very-low phosphate also tracks mortality (the U-shape) — undermined the case for aggressive lowering of mildly elevated values [8]. The combined effect: KDIGO acknowledged uncertainty, kept the directional priority, dropped the numerical target.

3.1 The observational signal is large and consistent

Higher serum phosphate is associated with mortality in dialysis patients across virtually every cohort that has examined the question. The headline data:

• Block GA 2004 (J Am Soc Nephrol) — the foundational 40,538-patient analysis from the Fresenius database. Serum phosphate > 5.0 mg/dL was associated with progressively increased all-cause mortality; the relationship was monotonic and steep above 6.5 mg/dL [2].
• DOPPS — across multiple international cohort waves (12 countries, > 25,000 hemodialysis patients), serum phosphate > 5.5 mg/dL was consistently associated with increased all-cause and cardiovascular mortality. The DOPPS analyses also identified the U-shape — phosphate < 3.5 mg/dL carried similar or greater mortality risk than mid-range values, plausibly reflecting protein-energy wasting and inadequate intake rather than the phosphate level itself [4].
• Recent UpToDate-cited meta-analysis — 12 studies, 92,345 patients (approximately 97% on dialysis): per 1 mg/dL increase in serum phosphorus, mortality risk increased by 18% (95% CI 1.12–1.25) [9].
• Nondialysis CKD meta-analysis — 3 studies, approximately 5,000 patients with non-dialysis CKD: 35% increase in mortality per 1 mg/dL increase (95% CI 1.16–1.57) [9].

The biological framework is mechanistically coherent. Elevated phosphate drives FGF-23 elevation, secondary hyperparathyroidism, vascular smooth muscle calcification (via Pit-1 transporter and osteoblastic transdifferentiation), and accelerated arteriosclerosis. The cardiovascular burden of CKD-MBD is a real biology, not a statistical artifact.

3.2 The U-shape

Phosphate behaves like albumin and BMI in the dialysis population. Very-high values track mortality through the calcification-FGF23-cardiovascular-event pathway. Very-low values track mortality through protein-energy wasting, malnutrition, inflammation, and inadequate intake. The U-shape implies that aggressive lowering of mildly-elevated phosphate may not be benign — particularly if it requires aggressive dietary protein restriction in patients already at risk for malnutrition [8,10].

This nuance is missed by quality metrics that score “percent of patients with phosphate in target range” without distinguishing the direction of out-of-range. A patient with phosphate 3.0 mg/dL on a low-protein diet may carry worse prognosis than a patient with phosphate 6.0 mg/dL adequately nourished and on appropriate binder therapy. The metric does not capture this.

3.3 Hard outcomes beyond mortality

• Vascular and valvular calcification — mechanistic and observational data consistently link phosphate to accelerated coronary artery calcium progression and cardiac valve calcification. The Treat-to-Goal trial (Chertow 2002) used coronary and aortic calcification as the primary endpoint specifically because the imaging signal was clear [11].
• Fracture — disordered mineral metabolism is associated with fracture risk in dialysis, but the relationship to serum phosphate specifically is confounded by PTH, vitamin D status, bone turnover state, and adynamic bone disease.
• Hospitalization — DOPPS and US Renal Data System analyses associate hyperphosphatemia with increased hospitalization, with mineral metabolism abnormalities clustering with cardiovascular admissions.

3.4 Causal inference — the limit of observational data

The observational signal is consistent and biologically plausible. It is also confounded by everything that elevates phosphate in the first place: residual kidney function, dialysis adequacy, dietary intake, adherence to binders, GI absorption variability, and the unmeasured comorbidity burden that comes with poor phosphate control. The patients who maintain phosphate < 5.5 mg/dL are also the patients more adherent to therapy, more nutritionally stable, and more clinically engaged. Mortality may track phosphate because both track an underlying gestalt of disease severity.

Mendelian randomization analyses have been attempted but the genetic instruments for phosphate are weak. The clean answer would be a randomized trial of phosphate target. We did not have one until very recently.

4.1 What we have

• DCOR (Suki 2007, Kidney Int) [6] — the largest randomized trial of phosphate-binder choice ever conducted. 2,103 incident hemodialysis patients randomized to sevelamer vs calcium-based binder, primary endpoint all-cause mortality. No overall mortality difference (HR 0.93, 95% CI 0.79–1.10). Subgroup analysis suggested benefit in patients age ≥ 65 (HR 0.77, p = 0.02), but the prespecified primary analysis was null.
• INDEPENDENT (Di Iorio 2012–2014) [7] — Italian incident-dialysis trial of sevelamer vs calcium carbonate. Reported a survival advantage for sevelamer, but trial size, attrition, and methodological concerns have limited its impact on guideline writing.
• Treat-to-Goal (Chertow 2002, Kidney Int) [11] — randomized 200 hemodialysis patients to sevelamer or calcium-based binder for 52 weeks. Primary endpoint: change in coronary artery calcium score and aortic calcification. Sevelamer arm showed less progression of coronary and aortic calcification. Surrogate endpoint, not mortality.
• HiLo Pragmatic Trial (Edmonston, Wolf, et al., design published 2021 — AJKD 77(6):920-930) [13] — multicenter, pragmatic, cluster-randomized trial enrolling 4,400 patients across 80–120 DaVita and University of Utah dialysis facilities (NCT04095039). Compares phosphate binder prescriptions and dietary recommendations to achieve a “Hi” serum phosphate target (≥6.5 mg/dL) versus a “Lo” serum phosphate target (<5.5 mg/dL). Primary outcome: hierarchical composite of all-cause mortality and all-cause hospitalization. As of the most recent indexed PubMed records, only the trial design and rationale are published; the trial result paper has not yet appeared in indexed literature. This is the field’s first adequately-powered randomized test of phosphate target effects on patient-important outcomes — its eventual publication will be the most consequential piece of evidence in this area in two decades. The pragmatic-trial-design framework underlying HiLo is described by Wald and colleagues [12]. Until results publish, the central skeptical question of whether lowering phosphate improves hard outcomes remains formally untested by RCT.

4.2 Surrogate vs patient-relevant outcomes

The binder-choice trials have consistently shown: - Surrogate improvement with non-calcium agents (less coronary artery calcium progression, lower FGF-23 levels, lower calcium burden) - No clear hard-outcome improvement in adequately-powered randomized comparisons

This is the central tension. Nephrology has spent two decades treating phosphate based on the assumption that the surrogate-outcome story would translate to hard outcomes. The trial evidence has not delivered. The pragmatic field response has been to keep treating phosphate (because the observational signal is real, the biology is plausible, and the regulatory and quality framework demands it) while acknowledging that the population-level mortality benefit of any specific intervention is unproven.

4.3 The honest assessment

Aggressive phosphate lowering in maintenance dialysis is guideline-by-consensus and observational-data-driven, not RCT-driven. The KDIGO 2017 weak recommendation language reflects this honestly. The pragmatic clinical position:

• Treat clearly elevated phosphate (e.g., > 6.5 mg/dL on serial measurements) — the observational signal is strong enough to act on
• Avoid the false precision of chasing a specific numerical target
• Recognize that pill burden, GI tolerability, adherence, and patient quality of life are real costs of binder therapy
• Be skeptical of any aggressive intervention that has not been shown to improve patient-relevant outcomes

5.1 Comparative table

5.2 Calcium vs non-calcium — the mortality question

The COSMOS observational analysis and several systematic reviews have found higher mortality with calcium-based binders compared with non-calcium binders [5,17]. The clearest randomized signal comes from INDEPENDENT (Di Iorio) [7]. DCOR did not replicate this overall (only in age ≥ 65 subgroup) [6]. KDIGO 2017 reflects the uncertainty by recommending restricting calcium-based binder dose rather than declaring sevelamer or other non-calcium agents superior.

The pragmatic position: avoid calcium-based binders in patients with known vascular calcification, low PTH, adynamic bone disease, or hypercalcemia. Use them when cost and pill burden tip the balance — generic calcium acetate remains the cheapest and is reasonable in selected patients.

5.3 Iron-based binders — the dual benefit

Ferric citrate (Auryxia) and sucroferric oxyhydroxide (Velphoro) are both effective phosphate binders with very different iron-absorption profiles. Ferric citrate is absorbed substantially as elemental iron — the Lewis 2015 pivotal trial demonstrated meaningful increases in serum ferritin and TSAT, with a corresponding reduction in IV iron and ESA requirements [14]. This is a real bundled-payment economic argument when dialysis units are managing IV iron and ESA expense. Sucroferric oxyhydroxide is minimally absorbed, so it does not produce the iron-stores benefit but offers the lowest pill burden in the field (approximately 3 tablets/day) — a meaningful adherence advantage [15].

5.4 Pill burden, adherence, and the patient experience

The average dialysis patient on phosphate binders takes 8–12 tablets per day, often spread across all meals, alongside the rest of their dialysis pharmacopeia. Pill burden is a primary driver of non-adherence, and non-adherence is the most common reason for “refractory” hyperphosphatemia. Any honest binder discussion must include patient-experience and adherence reality, not just biochemistry.

6.1 Mechanism

Tenapanor is a minimally-absorbed, gut-restricted inhibitor of intestinal sodium/hydrogen exchanger 3 (NHE3). NHE3 inhibition reduces intestinal sodium absorption, which in turn lowers intestinal pH and water content. The downstream consequence — and the basis for the phosphate effect — is a tightening of paracellular permeability for phosphate across the intestinal epithelium. Tenapanor does not bind phosphate. It reduces phosphate absorption by blocking the paracellular pathway that accounts for the majority of intestinal phosphate uptake [18].

This matters operationally: - No pill-with-every-meal binder dosing pattern — tenapanor is one 30 mg tablet twice daily, taken just before meals - No binder-bound phosphate competition with other GI medications - GI side effects driven by mechanism — softened stool and diarrhea result directly from increased intestinal sodium and water content. Roughly half of patients experience some change in bowel pattern, with discontinuation rates in clinical trials of 1.7–3.4% [20] for tenapanor and loosened stools in 53% [21] in the longer trial (largely manageable with dose adjustment).

6.2 The trial program — BLOCK, AMPLIFY, PHREEDOM

The trial program established (a) tenapanor as effective monotherapy in a binder-naïve population (BLOCK), (b) tenapanor as additive to binder therapy when binders alone are insufficient (AMPLIFY), and (c) durable efficacy through 52 weeks (PHREEDOM). The FDA approved tenapanor in October 2023 for hyperphosphatemia in adults with CKD on dialysis — initially as monotherapy, later expanded to include add-on use after the AMPLIFY data.

6.3 The diarrhea question — manageable but real

Loosened stools are a mechanism-based, not idiosyncratic, side effect. The clinical management strategies that work: - Start at the full dose — most patients adapt within 1–2 weeks - Dose-reduce to 30 mg once daily if loose stools persist past the adaptation window - Position before meals — taking immediately before eating optimizes GI transit timing - Counsel patients explicitly that softened stool is the mechanism working, not a “drug failure” - Discontinue and switch only if functional impairment occurs — for many patients, mild loose stool is acceptable in exchange for binder pill-burden reduction

6.4 Positioning — monotherapy, add-on, or binder replacement

The current operational positioning emerging from the trial data and post-marketing experience:

• Add-on to binders (AMPLIFY indication) — the most common real-world use. The patient is on adequate binder therapy but cannot achieve target phosphate without further intervention. Tenapanor adds a different mechanism to the existing approach. This avoids the binder-replacement decision entirely.
• Monotherapy (BLOCK indication) — appropriate for patients who genuinely cannot tolerate binders, who have severe pill burden problems, or who have had GI tract complications from calcium-based binders. The pill-count-reduction argument is real here.
• Long-term durability (PHREEDOM data) — supports continued use through 52 weeks without efficacy loss, which addresses the “transient effect” concern that some clinicians raised after BLOCK.

The patient phenotype that benefits most from tenapanor: - High pill burden with non-adherence - GI side effects from binders (constipation from calcium, GI intolerance from sevelamer, stool changes from iron-based agents) - Refractory hyperphosphatemia despite adequate binder therapy - Patient preference for fewer total tablets - Capacity to manage mild bowel pattern changes

6.5 Cost and coverage — Part B vs Part D, TDAPA, the bundle debate

This is the active operational and policy question for 2025–2026 nephrology practice.

The CMS ESRD bundled prospective payment system (PPS) historically excluded oral-only ESRD-related drugs from the bundle, with separate billing under Medicare Part D. Tenapanor was approved in 2023 and entered Medicare coverage initially through Part D. CMS proposed inclusion of oral-only ESRD drugs (including tenapanor and oral phosphate binders) in the ESRD bundle, with TDAPA (Transitional Drug Add-on Payment Adjustment) status as a transition mechanism.

The current 2025–2026 status: - Tenapanor is currently in TDAPA status for two years following bundle inclusion (timeline subject to CMS rule changes) - Phosphate binders themselves were also targeted for bundle inclusion under the 2025 ESRD PPS Final Rule, with TDAPA support. This is the operational change with the largest impact: dialysis units now receive bundled payment for binders that were previously billed separately under Part D - Practical effect — dialysis units now have direct economic exposure to binder choice in a way they did not before. Selection toward lower-cost binders (calcium carbonate, generic calcium acetate) becomes economically rational at the unit level even when clinical considerations argue for non-calcium agents - Tenapanor pricing is driven by Ardelyx (the manufacturer); during TDAPA coverage, dialysis units can prescribe without absorbing the full agent cost, but at TDAPA expiration the bundle calculation must absorb the agent cost or the unit will lose money on tenapanor prescriptions

The implication for practice: the next 2–3 years will determine whether tenapanor remains broadly accessible or becomes a payer-formulary-restricted agent based on bundle economics. Nephrologists who want to use tenapanor for the right patient should document the indication and rationale clearly to support coverage decisions.