Clinical reference

Anaemia and mineral bone disorder

Two areas where the guideline has spent fifteen years walking back its own confidence. In anaemia the ceiling turned out to matter more than the floor. In mineral bone disorder the fixed numeric targets turned out not to be supportable at all, and were replaced by trends.

This page gives principles and published targets. It gives no doses, for any agent, and it will not.

A threshold is not a dose. Knowing that KDIGO suggests initiating an erythropoiesis-stimulating agent in a dialysis patient at a haemoglobin of 9.0 to 10.0 g/dL or below tells you when the conversation starts. It tells you nothing about what to give, how much, how often, or how to titrate it, and those decisions belong to the treating team, the local formulary and the individual patient.

Anaemia: the published thresholds

Investigate and correct all correctable causes of anaemia, including iron deficiency, before starting an ESA or a HIF-PHI.

KDIGO Anemia in CKD, 2026

In dialysis patients, initiation of an ESA or HIF-PHI is suggested when the haemoglobin is 9.0 to 10.0 g/dL or below.

KDIGO Anemia in CKD, 2026

Maintenance target: haemoglobin below 11.5 g/dL, in both dialysis and non-dialysis CKD.

KDIGO Anemia in CKD, 2026

Do not intentionally exceed a haemoglobin of 13 g/dL. This is a long-standing KDIGO position, carried forward from 2012, and it rests on the harm signal in the normalisation trials: CHOIR, CREATE and TREAT.

KDIGO Anemia in CKD, 2026

Adopt a restrictive transfusion policy, particularly in transplant candidates, to minimise allosensitisation.

KDIGO Anemia in CKD, 2026

Where the guidelines disagree: the haemoglobin floor, and HIF-PHIs

KDIGO, 2026: Suggests initiating an ESA or HIF-PHI in dialysis patients at a haemoglobin of 9.0 to 10.0 g/dL or below, with a maintenance target below 11.5 g/dL. Recommends ESAs as first-line over HIF-PHIs (Grade 2D), on the grounds of established safety and accumulated experience.
ERBP commentary, 2026: Emphasises not routinely allowing haemoglobin to fall below 10 g/dL, and suggests a 10 to 12 g/dL range with individualisation. Argues that both ESAs and EMA-approved HIF-PHIs should be considered treatment options rather than placing ESAs first.

This divergence is real and current, and a clinician should know which position their unit protocol has adopted and why. ERBP additionally raises sex-specific ESA responsiveness, pregnancy-specific haemoglobin thresholds, the effect of SGLT2 inhibitors on erythropoiesis and iron metabolism, and iron correction in restless legs syndrome, none of which KDIGO addresses in the same detail.

Iron

In haemodialysis: initiate or continue intravenous iron when TSAT is 30% or below AND ferritin is 500 ng/mL or below.

KDIGO Anemia in CKD, 2026 (consistent with KDIGO 2012)

In non-dialysis CKD: ferritin below 100 ng/mL and TSAT below 40%, or ferritin 100 to 299 ng/mL and TSAT below 25%.

KDIGO Anemia in CKD, 2026

Iron ceiling: withhold routine iron if ferritin is 700 ng/mL or above, or TSAT is 40% or above.

KDIGO Anemia in CKD, 2026Practice point

Two caveats worth carrying into the clinic. Ferritin is an acute-phase reactant as well as a measure of iron stores, so in an inflamed dialysis population it can be misleadingly high, and a normal ferritin does not exclude functional iron deficiency. TSAT is the better guide to iron actually available for erythropoiesis.

And the iron ceiling is a practice point, not a graded recommendation. ERBP questions specifically whether the TSAT above 40% threshold is adequately evidence-based, and cautions against generalising the findings of PIVOTAL, which tested proactive high-dose intravenous iron in haemodialysis, to all CKD populations (ERBP commentary, Nephrology Dialysis Transplantation, 2026).

Monitoring

Haemoglobin, with the direction of travel mattering as much as the value. A rapidly rising haemoglobin on treatment is a reason to act, not to wait for it to cross a threshold.
Ferritin and TSAT together. Neither alone is sufficient, and ferritin alone in an inflamed patient is actively misleading.
Look for the correctable causes before escalating: iron deficiency, blood loss (including in the dialysis circuit and from repeated phlebotomy), B12 and folate, haemolysis, infection and inflammation, hyperparathyroidism, malignancy and haemoglobinopathy.
ESA hyporesponsiveness is a clinical signal, not a dosing problem. A patient who needs escalating doses to reach the same haemoglobin usually has something else going on, and escalating regardless is where the harm in the normalisation trials came from.

CKD-MBD: from targets to trends

The defining change in the 2017 CKD-MBD update was the move away from fixed numeric targets and towards serial trends and individualisation. That has not fully reached practice, and clinicians still routinely treat a single out-of-range phosphate result as though it were an instruction.

In CKD G3a to G5D, lower elevated phosphate levels toward the normal range. Treat progressively or persistently elevated phosphate, not a single high result.

KDIGO CKD-MBD update, 2017Grade 2C

Avoid prophylactic phosphate binder use in patients whose phosphate is normal.

KDIGO CKD-MBD update, 2017

Restrict the dose of calcium-based phosphate binders in adults with CKD G3a to G5D. This was strengthened from the 2009 wording.

KDIGO CKD-MBD update, 2017Grade 2B

Avoid hypercalcaemia in adults with CKD G3a to G5D. Maintain serum calcium in the normal range.

KDIGO CKD-MBD update, 2017Grade 2C

Use a dialysate calcium concentration of 1.25 to 1.50 mmol/L in CKD G5D.

KDIGO CKD-MBD update, 2017Grade 2C

In G5D requiring PTH-lowering treatment, calcimimetics, calcitriol, vitamin D analogues, or a combination, are all acceptable first-line options. No hierarchy is imposed.

KDIGO CKD-MBD update, 2017Grade 2B

In CKD G3a to G5 not on dialysis, calcitriol and vitamin D analogues are not routinely recommended. Reserve them for severe and progressive hyperparathyroidism. The optimal PTH level in non-dialysis CKD is not known.

KDIGO CKD-MBD update, 2017

Measure bone mineral density to assess fracture risk if the result will alter management. This was upgraded in 2017 from the previous position of not testing.

KDIGO CKD-MBD update, 2017Grade 2B

PTH, and why the target is a multiple rather than a number

In dialysis patients, the widely used target is an intact PTH of approximately 2 to 9 times the upper normal limit for the assay, with the instruction to treat marked changes in either direction within that range rather than reacting to a single value (KDIGO CKD-MBD, Grade 2C, carried forward from the 2009 guideline).

The reason the target is expressed as a multiple of the assay upper limit, rather than as an absolute number, is worth understanding rather than memorising: PTH assays are not interchangeable between laboratories. A number that means one thing on one platform means something materially different on another. If your unit has changed assay, your historical PTH trend has a discontinuity in it that no one may have flagged.

Vitamin D, analogues and binders

Nutritional vitamin D and active vitamin D analogues are two different interventions and are routinely conflated, including by prescribers.

Cholecalciferol and ergocalciferol are the vitamin, used to correct simple deficiency, and they require renal activation. Calcitriol, alfacalcidol, paricalcitol and doxercalciferol are pre-activated because the failing kidney cannot perform the final 1-alpha-hydroxylation step, and they are used to suppress an overactive parathyroid gland. They are not a vitamin top-up.

That distinction has a consequence: active vitamin D increases gut absorption of both calcium and phosphate, so overtreatment causes hypercalcaemia and worsens vascular calcification. It is why KDIGO does not recommend their routine use in non-dialysis CKD.

Phosphate binder principles

Binders act in the gut lumen, so they only work when they are taken with food, at every meal and with snacks. Timing failure is the commonest reason for an apparently binder-resistant phosphate, and it is usually an explanation problem rather than an adherence problem.
Calcium-based binders (calcium acetate, calcium carbonate) are effective and cheap, but add a calcium load. Guidance is to restrict their dose in G3a to G5D because of vascular calcification concerns.
Non-calcium binders include sevelamer, lanthanum carbonate, ferric citrate and sucroferric oxyhydroxide. Iron-based binders can also raise iron stores, which is relevant when you are simultaneously managing iron indices for anaemia.
Aluminium-based binders are avoided for long-term use because of aluminium toxicity, including bone disease and encephalopathy.
Pill burden is a clinical problem, not a complaint. Binders are frequently the largest component of a dialysis patient tablet count, and burden drives non-adherence, which then presents as treatment failure.
Never allow a phosphate problem to be solved by cutting protein. The highest-phosphate foods are often the highest-protein foods, and dialysis patients need protein. Phosphate additives in processed food are almost completely absorbed, while plant phosphate is bound as phytate and poorly absorbed, so the dietary target is additives, not lentils.

Class-level principles only. No agent on this list is recommended over another here, and no dose appears anywhere on this site.

Sources

KDIGO Clinical Practice Guideline for Anemia in CKD, 2026 (updating the 2012 anaemia guideline) (KDIGO, 2026)
European Renal Best Practice commentary on the KDIGO 2026 Anemia in CKD guideline (ERBP, Nephrology Dialysis Transplantation, 2026)
KDIGO 2017 Clinical Practice Guideline Update for the Diagnosis, Evaluation, Prevention and Treatment of CKD-MBD (KDIGO, 2017)
UK Renal Association commentary on the KDIGO 2017 CKD-MBD guideline update (UK Kidney Association, 2018)
PIVOTAL: proactive high-dose intravenous iron in haemodialysis (New England Journal of Medicine, 2019)
Phosphorus and your diet, and phosphate binders (National Kidney Foundation)