Erythropoietin
The kidney is the body's oxygen meter, and erythropoietin is the message it sends to the marrow — so when the kidney scars, nobody tells the marrow to make red cells.
What it normally does
Peritubular interstitial fibroblast-like cells in the renal cortex and outer medulla measure oxygen and make most of the body's erythropoietin (the liver makes a small amount, and is the main source in the fetus). When oxygen is low, the transcription factor HIF-2 alpha escapes destruction and switches on the erythropoietin gene. When oxygen is normal, prolyl hydroxylase enzymes (PHD1-3) hydroxylate HIF-2 alpha, the von Hippel-Lindau protein then ubiquitinates it and the proteasome destroys it within minutes (the HIF-PHD-VHL oxygen-sensing axis). These interstitial cells sit exactly where the kidney scars.
Explains why chronic kidney disease causes anaemia with no blood loss, why VHL disease and VHL-mutant tumours can drive erythropoietin production, and why a drug that blocks prolyl hydroxylase can restart the whole system.
Erythropoietin travels to the bone marrow and binds its receptor on committed red cell precursors (late BFU-E, CFU-E and proerythroblasts). Its dominant action is survival — it rescues precursors from apoptosis — and it also drives their proliferation and differentiation. Either way it cannot act faster than erythropoiesis itself: reticulocytes take days and a measurable haemoglobin rise takes weeks.
Explains why erythropoietin drugs correct anaemia over weeks, not hours, so they are useless for a patient bleeding today.
Erythropoietin achieves little without iron to build haemoglobin with. Iron leaves gut enterocytes and macrophages through ferroportin, the only cellular iron exporter so far identified. The liver hormone hepcidin binds ferroportin, occluding it and triggering its internalisation and degradation. Hepcidin rises with inflammation (IL-6) and with iron loading, and it is a small peptide cleared by the kidney, so it climbs in kidney disease.
Explains why patients on erythropoietin stop responding, why their ferritin looks normal or high while they are functionally iron starved, and why oral iron so often fails.
Haematocrit is the dominant determinant of whole-blood viscosity, so anaemia gives thin blood; that, plus hypoxic vasodilation, lowers systemic vascular resistance and produces the hyperdynamic circulation of chronic anaemia. Raising the haemoglobin reverses this. Erythropoietin also acts directly on the vasculature — raising endothelin-1, increasing sensitivity to angiotensin II and noradrenaline, and promoting vascular remodelling — so an ESA can raise blood pressure before the haematocrit has moved far.
Explains why pushing the haemoglobin back to normal with a drug raises blood pressure and promotes thrombosis, and why the blood pressure rise is not explained by viscosity alone.
What goes wrong
- Anaemia of chronic kidney disease← from “Peritubular interstitial fibroblast-like cells…”
Progressive tubulointerstitial fibrosis damages the oxygen-sensing interstitial cells — they transdifferentiate into myofibroblasts and largely stop making erythropoietin — so the signal fails even though the sensed hypoxia is real. Uraemia also shortens red cell survival and blunts the marrow response. The marrow itself is structurally normal, so the cells that are made are of normal size and colour; there are just too few. Anaemia becomes progressively more common as eGFR falls below about 60 mL/min/1.73 m2, affects roughly half of patients by an eGFR of 30, and is near-universal at end-stage. It is largely a diagnosis of exclusion: iron deficiency, B12/folate deficiency, occult blood loss, haemolysis and myeloma should be looked for first.
Normocytic normochromic anaemia + reduced eGFR + low reticulocytes, with other causes excluded = anaemia of CKD. About one in ten Australian adults has biomedical signs of CKD, and treated end-stage kidney disease is several times more common in Aboriginal and Torres Strait Islander people than in other Australians — many times higher again in remote central and northern Australia — so this anaemia is met early and often in those communities.
You would find: Tired, pale, breathless on stairs. Full blood count shows a normocytic normochromic anaemia with low reticulocytes in someone with reduced eGFR. The erythropoietin level (rarely measured in practice) is low or normal — inappropriately normal for that haemoglobin. Iron studies, B12 and folate are unremarkable.
- Functional iron deficiency and erythropoietin resistance← from “Erythropoietin achieves little without iron to…”
Kidney disease is an inflammatory state and hepcidin is cleared by the kidney, so hepcidin is high. High hepcidin shuts ferroportin, so iron cannot leave enterocytes or macrophages. Total body iron may be plentiful but it is locked away, so the marrow starves while ferritin — an acute phase protein as well as an iron store — reads normal or high. Give erythropoietin into that and little happens.
Before blaming the erythropoietin dose, check iron. Low transferrin saturation with normal or high ferritin is functional iron deficiency, and it is among the commonest reversible reasons an erythropoiesis-stimulating agent stops working.
You would find: Haemoglobin refuses to rise despite escalating erythropoietin doses. Iron studies show transferrin saturation under 20% with a normal or raised ferritin: iron present but undeliverable. Oral iron has been tried and failed.
- Over-correction: hypertension and thrombosis← from “Haematocrit is the dominant determinant of who…”
Raising the haemoglobin raises the haematocrit and blood viscosity and reverses the low resistance the patient had adapted to, while erythropoietin acts directly on vessels (endothelin-1, enhanced vasoconstrictor sensitivity). Peripheral resistance climbs, blood pressure climbs, and thickened blood thromboses where flow is slow. Trials targeting normal or near-normal haemoglobin (Normal Hematocrit, CHOIR, CREATE, TREAT) showed no benefit, and several showed harm — TREAT roughly doubled stroke (about 5% versus 2.6%). Part of the harm appears to track with the high ESA doses needed in hyporesponsive patients rather than with the haemoglobin itself.
The goal is to relieve symptoms and avoid transfusion, not to normalise the haemoglobin. Chasing a normal haemoglobin with an ESA increases stroke and death.
You would find: Blood pressure creeping up over weeks on an erythropoiesis-stimulating agent; a dialysis fistula or graft that keeps clotting; at the extreme, headache, seizures and hypertensive encephalopathy. Hence an Australian treatment target of roughly 100-115 g/L — deliberately below normal.
- Secondary polycythaemia (too much erythropoietin)← from “Peritubular interstitial fibroblast-like cells…”
Two routes. Either the sensor is doing its job in a genuinely hypoxic patient — severe COPD, untreated obstructive sleep apnoea, heavy smoking, a right-to-left cardiac shunt, altitude — or a tumour makes erythropoietin autonomously: renal cell carcinoma, hepatocellular carcinoma, cerebellar haemangioblastoma, phaeochromocytoma, uterine leiomyoma. Of these, haemangioblastoma, clear cell renal cell carcinoma and phaeochromocytoma are the ones that belong to von Hippel-Lindau disease, where the HIF pathway is disinhibited (hepatocellular carcinoma is not a VHL tumour). Either way the marrow is driven hard and red cell mass rises. Contrast this with apparent (relative) polycythaemia from plasma volume depletion, where red cell mass is normal.
High haemoglobin: measure erythropoietin. Low EPO with JAK2 V617F = polycythaemia vera. High or normal EPO = find the hypoxia, or image the kidneys and liver for the tumour.
You would find: High haemoglobin and haematocrit, often incidental, sometimes with headache or a plethoric face. The erythropoietin level splits the diagnosis: raised or inappropriately normal points to hypoxia or an erythropoietin-secreting tumour; suppressed points to an autonomous marrow — polycythaemia vera, where aquagenic pruritus (itch after a hot shower), splenomegaly and JAK2 V617F belong.
What we give, and how it works
Open a drug to see what it binds, what that does to the cell, and what you then see in the patient.
- Binds
- The erythropoietin receptor on committed red cell precursors in the bone marrow.
- Which does
- The receptor already sits as a preformed dimer; erythropoietin binding reorients it so the associated JAK2 kinases activate, switching on STAT5 and raising anti-apoptotic proteins such as Bcl-xL. Precursors that would have died survive, proliferate and mature.
- So you see
- Reticulocytes rise within about a week, haemoglobin over roughly two to six weeks. Fatigue lifts and transfusions become unnecessary. Darbepoetin alfa and the pegylated agent are engineered for a longer half-life, so they are given less often. (Epoetin alfa is the name used in most overseas literature; the plain epoetins marketed in Australia are epoetin lambda and epoetin beta.)
- And the same mechanism causes
- The rising red cell mass that fixes the anaemia also thickens the blood, and erythropoietin acts on vessels directly — so blood pressure climbs and clots form where flow is slow, most visibly a dialysis access that keeps blocking, and at the extreme stroke or hypertensive encephalopathy. The side effect is the therapeutic effect taken too far. Rarely, antibodies raised against the injected epoetin cross-react with native erythropoietin and cause pure red cell aplasia — a sudden loss of response with absent reticulocytes.
- Handling
- Iron status is corrected before starting or escalating, and blood pressure and haemoglobin are monitored regularly. Australian practice aims at a haemoglobin of about 100-115 g/L rather than a normal value, and a poor response prompts a search for the cause (iron, infection, inflammation, bleeding, hyperparathyroidism) rather than simple dose escalation.
Catches people out: ESAs shortened survival in randomised trials in several cancers (breast, head and neck, non-small cell lung), so they are avoided in patients receiving chemotherapy with curative intent. The mechanism is not settled — thrombosis is likely, and the older claim that tumour cells carry functional erythropoietin receptors rests on non-specific antibodies and is not established.
- Binds
- No pharmacological receptor. Given IV, the iron-carbohydrate complex is taken up by reticuloendothelial macrophages and released to transferrin (a small fraction is donated to transferrin directly).
- Which does
- Transferrin carries iron to erythroblasts via transferrin receptor 1, and they insert it into protoporphyrin IX to make haem. The marrow can finally build haemoglobin for the precursors the ESA rescued.
- So you see
- Haemoglobin rises, transferrin saturation climbs above 20%, and ESA requirements fall.
- And the same mechanism causes
- IV: delivering iron as a carbohydrate complex means some is handled as foreign particulate — hence infusion reactions, from the self-limiting Fishbane reaction (flushing, chest or back tightness) to rare anaphylaxis, and long-lasting brown skin staining if it extravasates. Ferric carboxymaltose in particular drives up FGF23, causing renal phosphate wasting and hypophosphataemia. Oral: in kidney disease much of the dose never leaves the bowel because hepcidin has shut the exporter, and that unabsorbed luminal iron is what blackens the stool and irritates the gut — the reason the tablet fails is the reason it upsets the gut.
- Handling
- Phosphate is worth checking in a patient on repeated ferric carboxymaltose who develops bone pain or proximal weakness. Because each dose of oral iron raises hepcidin for around a day, alternate-day dosing is absorbed better than several doses a day. Iron studies are unreliable in the days after an IV infusion.
Catches people out: Iron during active infection needs thought — the high hepcidin of inflammation exists partly to hide iron from pathogens, and IV iron is generally withheld during active bacteraemia.
- Binds
- The prolyl hydroxylase enzymes (PHD1-3) that hydroxylate HIF-2 alpha and mark it for VHL-mediated destruction when oxygen is plentiful.
- Which does
- With the enzyme inhibited, HIF-2 alpha survives and transcribes its targets: the erythropoietin gene, iron transport proteins, and less hepcidin. The cell behaves as though it were hypoxic when it is not.
- So you see
- The patient's own erythropoietin rises to something nearer physiological than the pharmacological peaks of an injected ESA, iron absorption and mobilisation improve, and haemoglobin rises — from a tablet. Trials show haemoglobin control comparable to ESAs, not superior.
- And the same mechanism causes
- HIF switches on the whole hypoxia programme, not just erythropoietin, and VEGF is part of that programme. Thrombosis (venous thromboembolism, vascular access clotting) is the consistent trial signal, with theoretical concern about driving diabetic retinopathy and tumour angiogenesis because those tissues answer to the same switch.
- Handling
- Blood pressure and haemoglobin are monitored as they are for an ESA — the viscosity and thrombosis problem does not disappear because the drug is oral.
Catches people out: Elegant on paper, but the safety story is still being written. Do not offer it in an exam as first-line Australian therapy.
- Binds
- No pharmacological target. This is replacement of the end product, not receptor pharmacology.
- Which does
- Donor red cells circulate immediately and carry oxygen.
- So you see
- Haemoglobin rises within hours and symptoms often settle the same day.
- And the same mechanism causes
- Donor cells carry foreign HLA and red cell antigens, so the recipient may make antibodies against them (alloimmunisation) — and in a CKD patient HLA sensitisation shrinks the pool of donor kidneys they can later be matched to and lengthens the wait. The foreignness that makes transfusion instant is what can cost them a transplant. Volume loading also matters in oliguric patients.
- Handling
- In a potential transplant recipient, transfusion is kept to genuine indications and the smallest number of units that will do, with reassessment after each unit.
Catches people out: Repeated transfusion also loads iron into liver and heart, which is why chronic transfusion is not a substitute for an ESA.
One hormone, two directions. Haemoglobin LOW with reduced eGFR and a normocytic normochromic film: anaemia of CKD once other causes are excluded — and check iron before blaming erythropoietin, because transferrin saturation under 20% with a normal or high ferritin is a common reason an ESA has stopped working. Haemoglobin HIGH: measure erythropoietin — suppressed means polycythaemia vera (JAK2 V617F), raised or inappropriately normal means hypoxia or an erythropoietin-secreting tumour.
Now test whether it stuck
Reading this through is not the same as being able to reconstruct it. Every question in the bank is free, with a full debrief on each option.