Intravenous Iron

Iron deficiency, the commonest cause of anaemia, is treated by iron supplementation. Oral iron, introduced by the French physician Pierre Blaud in the 19th century is the mainstay of therapy of iron deficiency. Oral iron is inexpensive and safe but is poorly absorbed and causes abdominal adverse effects in 35-59% of the patients. Injectable iron preparations were initially developed for the treatment of iron deficiency in patients intolerant to iron. Their use became more prevalent when it became clear that iron deficiency was a common cause of failure of erythropoiesis stimulating agent therapy and that oral iron was insufficient for this indication. The acceptance of intravenous iron was accelerate by the improved safety profile of the recently introduced parental iron preparation. The threshold for opting for intravenous iron is much lower than it was about two decades ago.

Is it possible to inject an iron salt, say ferric chloride, for iron deficiency, like it is to inject calcium chloride for hypocalcaemia? Iron, unlike calcium or potassium,  produces free radicals (oxyradicals) that can damage macromolecules and result in cellular injury. Ferric hydroxide,  the first injectable iron preparation used, caused an immediate release of iron in circulation resulting in severe reactions.  When an iron oxide, hydroxide or an iron salt is used, only a small dose can be safely administered. One estimate calculated the maximum iron permissible as 8mg/day (this is equivalent of the unbound iron binding capacity of transferrin). It would take about 6 months to bring up haemoglobin at this dose.

About  25mg of iron is delivered each day to the erythroid precursors for haemoglobin synthesis. Iron is a poorly absorbed micronutrient and stores are needed to  provide for sudden increase in demand as may be seen in patients with acute blood loss. Given the propensity of iron to cause free radical induced cellular injury, transport and storage systems capable of protecting the body from iron have evolved. Iron is transported bound to transferrin and stored as ferritin and haemosidin. In both the situations iron is bound to apoproteins (apotrasferrin and apoferritin) and this binding does not allow free radical generation.

Erythrocyte destruction takes place in the macrophages and macrophages have systems for the handling, storage and recycling of iron. Injectable iron preparations have a carbohydrate shell that prevents exposure of the plasma to free iron.  Injectable iron preparations are taken up by macrophages of the reticuloendothelial system by endocytosis. The iron is released and enters the macrophages iron pool. This iron has a fate similar to iron reaching the macrophage from other sources viz. intestinal absorption and erythrocyte destruction. Binding iron to carbohydrate shell has made it possible to administer as high as dose as 1000mg over as short a time as 15 minutes. This is a 125 fold increase over the amount of iron that can be administered without the carbohydrate shell.

Injectable iron preparations share a common structure. They have a core of iron oxide/hydroxide that is associated with a carbohydrate shell. The carbohydrate shell  keeps free iron from entering the plasma and in a way performs the same function as transferrin. The antigenicity of the shell and the strength of binding between the iron core and the carbohydrate shell determines the side effects and the maximum dose per injection.

Preparations of Injectable Iron

The preperations of injectable iron include

  1. Preparations with a strong association between iron core and the carbohydrate shell
    1. High-Molecular weight iron dextran
    2. Low -molecular weight iron dextran
    3. Ferric carboxymaltose
    4. Ferumoxytol
  2. Perperations with a weak association between iron core and carbohydrate shell
    1. Iron Sucrose
  3. Preperations that have a labile low molecular weight components
    1. Ferric Gluconate
    2. Iron-Sorbitol-Citric Acid Complex, Dextrin-Stabilized
Table 1 – parentral iron preparations
Drug Shell MW T1/2 Labile
Dose and administration
HMW ID High moleculer weight dextran 265 60 1-2% Should be administered 1 hr after an intravenous test dose (0.5ml over at least 5 minutes) Maximum dose 20mg/kg. Administered as intravenous bolus 100mg/day slowly. A total dose infusion may be given over a prolonged period.
LMW ID  Low molecular weight dextan  165  20  1-2% Should be administered 1 hr after an intravenous test dose (0.5ml over at least 30 seconds). It may be administered as daily iv bolus of 100mg over 2 mins or a a total dose infusion over 3-4 hours may be administered 20mg/kg
Iron Sucrose  Sucrose 30-60  6  4-5% up to 300mg in a single dose, higher doses have been administered over as a prolonged infusion. May be administered undiluted as an iv bolus 200mg over at least 10 mins or as an infusion in 0.9% saline over 15-60 minutes
Ferric Gluconate  Gluconate  289-440  1  5-6% 125mg, iv bolus at 12.5mg/min or as a 1 hr infusion in 0.9% saline
Ferric Carboxy-maltose Carboxy-maltose  150  16  1-2% The maximum dose is 20mg/kg (max 1000mg). Doses of 200-500mg should be administered at 100mg/min, doses between 500mg and 1000mg should be administered over 15 minutes. It may be administered as an iv infusion diluted in 0.9% saline
Ferric Isomaltoside Isomaltoside 150  20  <1% May be administered as an intravenous bolus dose of 500mg un to three times a week at the rate of 50mg/min or 20mg/kg (maximum 1000mg) in 0.9% saline over 1 hour
Ferumoxytol Carboxy-methyl Dextran 750 15 <1% 510mg as a iv bolus at the rate of 30mg/sec. A second dose may be administered after 3-8 days

The parental iron preparations differ in the carbohydrate that surrounds the iron core. The therapeutic implications of these differences are:

  1. Antigenicity: Dextran is antigenic. Patinets may have performed antibodies to dextran or may develop antibodies during therapy. All iron preparations  containing dextran can give rise to anaphylaxis. The risk is greatest with high molecular weight iron dextran. Anaphylaxis may also be seen with low molecular weight iron dextran though the risk is lower and has also been reported with ferumoxytol that contains carboxymethyl dextran.
  2. Maximum dose per injection: All intravenous iron preparations tend to release labile iron. There is an inverse relationship between the amount of labile iron released and the strength of association between the iron core and carbohydrate shell. Higher the molecular weight of the iron preparation, stronger is the association. Labile iron is taken up by transferrin. Non-transferrin bound iron (NTBI) appears when the binding capacity of transferrin is overwhelmed. NTBI is responsible for tissue damage and restricts the dose of a injectable preparation that can be given in a single infusion. Iron preparations may be classified on the basis of strength of the associations between iron core and the carbohydrate shell. Iron dextran, ferric carboxymaltose and ferumoxytol, that have a high molecular weights have a tight binding between iron core and the carbohydrate shell, release up to 2% labile iron. These can be given in a large single dose and are suitable for total dose infusion. Iron sucrose has a weaker association between iron core and carbohydrate shell and has 4-5% labile iron. This limits the amount of drug that can be administered at one time to 300mg. Ferric gluconate is a large molecule but has two types of polymers. Lower weight polymers (about 18,000) have a weaker association between iron and carbohydrate limiting the dose to 125mg per dose.

Dose and Administration

The total dose of parenteral iron is calculated as follows

Total iron dose (mg) = weight (kg) X Hemoglobin deficit  X 0.24 + 500

Haemoglobin deficit (g/L) = Target Haemoglobin (g/L) – Actual Haemoglobin (g/L)

Adverse Effects

  1. Infusion Related Events: the risk of infusion related events (per million) is 0.6 for iron sucrose, 0.9 for ferric gluconate, 3.3 for low molecular weight dextran iron and 11.3 for high molecular weight iron dextran.  Iron sucrose has been safely administered to patients who having sensitivity to iron dextran.  Despite the fact that some studies suggest that low molecular weight iron dextran in as safe as iron sucrose, a test dose must be administered before any iron dextran preparation.
  2. Delayed Reactions to Intravenous Iron: A syndrome charecterized by fever, arthralgia and lymphadenopathy may be seen as a delayed consequence of intravenous iron therapy. Premedication with steroids may reduce the risk of this manifestation.

Further Reading

  1. Rodolfo Delfini Cançado, Manuel Muñoz. Intarvenous iron therapy: How far have we come? Rev Bras Hematol Hemoter. 2011; 33(6): 461–469.
  2. Hayat A. Safety Issues With Intravenous Iron Products in the Management of Anemia in Chronic Kidney Disease. Clin Med Res. Dec 2008; 6(3-4): 93–102.
  3. Peter Geisser and Susanna Burckhardt The Pharmacokinetics and Pharmacodynamics of Iron Preparations. Pharmaceutics. Mar 2011; 3(1): 12–33.



Etiology of Polycythaemia

The terms polycythemia and erythrocytosis though used interchangeably describe different entities. Polycythemia (Gk, polys + kytos, cell, haima, blood) is an absolute increase in the mass of red cells. Erythrocytosis refers to only increase in the concentration of erythrocytes. Erythrocytosis may be due to increase in red cells (polycythemia, absolute erythrocytosis) or decrease in plasma volume (relative erythrocytosis).

Erythropoiesis and It’s Control

Erythrocyte are produced from haemopoietic stem cells (see Morphology of Erythroid Processors). Erythropoietin (EPO) prevents apoptosis of erythroid precurssors and increases erythrocyte production. Hypoxia stimulates erythropoietin production.

Erythropoietin is produced in response to hypoxia. Hypoxia sensing involves

  • von-Hipple Lindau oncogene (VHL) protein
  • Hypoxia inducible factor (HIF) of there are three forms HIF1, HIF2 and HIF3. HIF2 is involved in erythropoietin production in response to hypoxia.
  • Prolyl-4-hydroxylase

Oxygen is critical for cell survival and rapid response to hypoxia is critical. Synthesizing proteins as effectors of response is slow process that may not be appropriate for a stress like hypoxia. Instead of synthesizing HIF in response to hypoxia the cells control the levels HIF by controlling it’s destruction. Cells on one hand synthesize HIF at a steady rate and on the other carry out a proteosomal degradation HIF by oxygen dependent mechanism.

Prolyl hydroxylase (PDH) is a constutively expressed enzyme that is involved in the process of marking HIF for proteosome mediated destruction. Proteosomal destruction of HIF also requires the von Hipple-Lindau oncogene protein. PHD is oxygen sensitive. Hypoxia reduces the PHD activity which in turn increases HIF levels by reducing its degradation. HIF induces the transcription of many genes involved in response to hypoxia including the erythropoietin gene. Increased HIF causes erythrocytosis. Mutations of the PDH and VHL gene have an effect similar to hypoxia. They increase erythrpoietin secretion and cause increased erythrocyte production.

Erythropoietin action is mediated by the erythropoietin receptor (ER). Binding of EPO to ER results in conformational change and the receptor is switched on. Receptors mediate their actions by a cascade of enzymatic actions. One of the most common mechanism of receptors action is activation of receptor tyrosine kinase activity. This allows the receptor to phosphorylate proteins and start a cascade of changes, usually phosphorylations, that lead to target actions. Not all receptors have tyrosine kinase activity. Those without tyrosine kinase activity rely of non-receptor tyrosine kinase for action. The erythropoietin receptor lacks tyrosine kinase activity essential for signal transduction. It associated with the non-receptor tyrosine kinase Janus kinase 2 (JAK2). JAK2 mediates the action of erythropoietin receptor (see ErythropoietinReceptor Signalling for details of erythropoietin production and action).

Table 1 – Causes of Polycythaemia
Primary Polycythaemia (serum erythropoietin low)
Inherited/Congenital Acquired
Familial Erythrocytosis Type I Polycythaemia Vera
Secondary Polycythaemia (serum erythropoietin normal/high)
Inherited/Congenital Acquired
  1. Disorders of Oxygen Sensing
    • Familial Erythrocytosis Type II (VHL Mutations)
    • Familial Erythrocytosis Type III (PHD2 mutations)
    • Familial Erythrocytosis Type II (HIF2A Mutations)
  2. High Affinity Haemoglobins
  3. Inherited Methaemoglobinaemia
  4. Congenital heart disease with right to left shunt
  1. Hypoxia
    • Chronic lung disease
    • Conginital heart disease with reversal of left to right shunt
    • Smokers
    • High altitude
  2. Ectopic erythropoetin seceretion
  3. Paraneoplastic Erythropoietin seceretion

Causes of Polycythemia (See Table 1)

Erythrocytosis may be primary or secondary, inherited or acquired. Primary erythrocytosis is erythropoietin independent. These patinets have low EPO levels. Patients with secondary erythropoiesis have have normal or high erythropoietin. The genetic defects associated with polycythaemia are listed in table 2.

  1. Primary Polycythaemia (Polycythaemia with low serum erythropoietin)
    1. Inherited
      1. Familial erythrocytosis type I: Inherited erythrocytosis is an autosomal dominant disorder resulting from truncating mutation in the erythropoietin receptor that result in deletion of between 57 to 127 amino acids from the cytoplasmic domain of the receptor (see table 2 below). The cytoplasmic domain appears to be necessary for inhibition of receptor action. Truncation results in the loss of the SHP-1 binding site. SHP-1 dephosphorylates JAK2 and attenuates erythropoietin rceptor signaling. Deletion causing familial erythrocytosis type I result in a loss of cytoplasmic portion of the erythropoietin receptor resulting in a receptor that can not interact with SHP-1 and be the turned off. Pathways stimulated by the erythropoietin receptor are continuously activated and polycythaemia is seen. The erythropoietin levels are low.
    2. Accquired
      1. Polycythaemia Vera: Polycythaemia vera is an accquired polycythaemic myeloproliferative disease. Majority (95%) of the patients have the JAK2V617F mutation. Patients not having JAK2V617F mutation have mutations in exon 12. JKAK2 is the mediator of effects of erythropoietin receptor. These mutations result in increased erythropoietin sensitiviy and polycythaemia (see table 2 below). The erythropoietin levels are low.
  2. Secondary Polycythaemia (Polycythaemia with normal or high serum erythropoietin
    1. Inherited
      1. Familial Erythrocytosis type 2 (Chuvash and other Polycythemias): Patients with familial erythrocytosis from the Chuvash republic of the former Soviet Union was found to have an Arg to Trp substitution at the amino acid position 200 (ARG200TRP) in the VHL gene. Other VHL mutations associated with erythrocytosis including VAL130LEU, ASP126TYR, PRO192SER and HIS191SER (see table 2 below). These patients have an abnormality in hypoxia sensing and result in an inappropriately high production for the degree of hypoxia. VHL muttions result in von Hipple-Lindau syndrome that predisposes the individula to renal cell carcinoma. No such predisposition is found in patients with type 2 familial erythrocytosis. Patients with Chuvash polycythaemia develop vertibral angiomas, are predisposed to venous thrombosis and respond abnormally to hypoxia
      2. Familial Erythrocytosis Type 3Familial erythrocytosis type 3 results from mutations of the PDH2 gene resulting in a defect in oxygen sensing (see table 2 below). These patients have inappropriately normal erythropoietin levels. Identified mutations are listed in table above.
      3. Familial Erythrocytosis Type 4: Familial erythrocytosis type 4 is an autosomal dominant disorder resulting from gain of function mutations in the HIF2A gene that codes for the α subunit of hypixia inducible factor 2 (see table 2 below). The mutations cause a defect in oxygen sensing resulting in an inappropriately high production of erythropoietin for the oxygenation.
      4. Haemoglobin Anomalies Causing Polycythaemia: Polycythaemia is caused by high affinity haemoglobins and congenital methaemoglobinaemia (see table 2 below). Mutations resulting in high affinity haemoglobins result in impaired oxygen release. this is sensed as hypoxia, erythropoietin production increased and polycythaemia results. Congenital methaemoglobinca results in suboptimal oxygen delivery erythropoietin secretion and polycythaemia.
    2. Acquired polycythaemia
      1. Accquired secondary polycythaemia: Accquired secondary polycythemia is observed in the following conditions
        1. Conditions associated with hypoxia:
          • Chronic Obstructive pulmonary disease
          • Left to right shunt with eisenmengerization
          • Chronic Smokers
          • High altitude polycythaemia
        2. Conditions associated with ectopic erythropoietin secretion:
          • Solitary renal cysts
          • Polycystic kidney disease
          • Hydronephrosis
        3. Paraneoplastic erythropoietin secretion:
          • Renal cell carcinoma – most common
          • Hepatoma – Less common
          • Occasional associations: Adrenal tumours,  cerebellar hemangioblastoma, hemangiomas, pheochromocytomas, sarcomas, uterine fibroids and Wilms tumor.
Table 2 – Genetic Defects Causing Polycythaemia
Syndrome Gene Mutations
Inherited disorders
Familial Erythrocytosis Type I (Autosomal Dominant) Erythropoietin Receptor
  • TRP439TER
  • 1-BP INS, 5975G
  • ASN487SER
  • 7-BP DEL, NT5980
  • 1-BP INS, 5967T
  • TYR426TER
  • 5968_5975DUP
Familial Erythrocytosis Type II (Autosomal reccssive) VHL Gene
  • ARG200TRP
    (Chuvash Polycythaemia)
  • VAL130LEU
  • ASP126TYR
  • PRO192SER
  • HIS191ASP
Familail Erythrocytosis Type III (Autosomal Dominant, all patinets have been heterozygous)  PDH2 gene
  • PRO317ARG
  • ARG371HIS
  • HIS374ARG;
Familial Erythrocytosis Type IV (Autosomal DOminant) HIF2A gene
  • GLY537TRP
  • GLY537TRP
  • MET535VAL
High Affinity Haemoglobin (Autosomal Dominant) HBA, HBB 93 varients listed at accessesd on 11th July 2013
Congenital Methemoglobinaemia Cytochrome b5Cytochrom b5 Reductase, Hemoglobin M The specific mutations are described in the links to the genes in the adjacent columns
2,3 Bisphosphoglucerate Mutase deficiency 2,3-Bisphoglyceromutase ARG89CYS; 1-BP DEL, 205C
Polycythaemia Vera JAK2  VAL617ILE (V617F), Exon 12 mutations

Myelocyte and Neutrophils – Band and Segmented

Myelo Band Neutro Final

Neutrophilic Cells. The images shows myelocyte (1), band neutrophils (2), two lobed neutrophil (3) and neutrophils (4)

The image about shows morphology of cells of the neutrophilic series. The first cell to show commitment to a particular granulcoytic series (eosinophilic, basophilic or eosinophilic) is the promyelocyte. The promyelocytes of the three series can however only be differentiated by electron microscopy. The earliest cell showing features of neutrophilic differentiation on staining with Romanowsky staining is the myelocyte. Myelocytes may be neutrophilic, eosinophilic and basophilic. Myelocyte is a round cell with a round to oval nucleus that may be eccentrically placed. The cytoplasm shows two types of granules.

  1. Primary granules: The primary granules are azurophilic (reddish-purple or burgundy color) and are remnants of the granules of  the promyelocyte stage. The myelocyte does not synthesize primary granules. As the cell matures the number of primary granules declines and they disappear by the cells matures to a polymorphonuclear neutrophil.
  2. Secondary granules: The the size and staining of secondary granules are specific to the type of granulocytes. In neutrophils these granules are fine and pink staining

As the myelocyte matures the nucleus becomes more indented and finally becomes lobed.

The image above captures these changes

  1. Cells labels (1) are myelocytes. These show a pink cytoplasm with few primary (azurophilic) and many secondary (fine pink) granules and have a oval eccentrically placed nucleus with clumped chromatin without a nucleolus.
  2. The myelocyte develops and indentation of the nucleus and matures to a metamyelocyte. The metamyelocyte matures to a band neutrophil (cells labeled 2). which is called so because the nucleus is band shaped. With maturation the nucleus develops lobulation. The differentiation between a band neutrophil and a two lobed neutrophil is arbitrary and of little practical importance. A band cell becomes a two lobed neutrophil when its nucleus develops a constriction that is more than half to two third of the nuclear width (cell labeled 3).
  3. A neutrophil usually has 2-5 nuclear lobes (cells labeled 4)

Normoblast Maturation

IMG_0354 Maturing Normoblast

The image above shows three polychromatophilic normoblasts. The one on the left is the least mature and the one on the right the most mature. Maturation is associated with

  1. Decrease in the size of the cell and nucleus: This is obvious.
  2. Clumping of chromatin: This is obvious.
  3. Increasing cytoplasmic acidophilia and decreased basophilia: The cells are flanked on each side by erythrocytes. The cytoplasm of the least mature cell (left most cell) is basophilic compared to that of the adjacent erythrocyte. The cytoplasm of the most mature cell (the right most cello) has lost basophilia and is almost the colour of the cytoplasm of the erythrocyte. 

Orthochromatophilic Normoblast


Orthochromatophilic normoblast

An orthochromatophilic normoblast is the most mature of nut;elated red cell precursors. The cytoplasm has lost almost all it’s basophilia and is the same colour as that of the erythrocyte. The nucleus is small pyknotic with densely condensed chromatin.

Embryonic Haemoglobins

Embryonic haemglobins are haemoglobins produced in the yolk sac stage of erythropoiesis. There are three embryonic haemoglobins Gower I, Gower II and Portland. Embryonic haemoglobins like their adult counterparts are tetramers made of two α-like (α and ζ)chains and two β-like chains(γ and β). The α-chain is a part of haemoglobins from the embryonic stage and complete loss of α-globin leads to anaemia from early gestation. This not the case with the β chain.

Haemoglobin is a tetramer of two α-like and two β-like globin chains. The genes of the α-like globins are found on chromosome 16 (16p13.3) and those for the β-like genes are found on chromosome 11 (11p15.5). The α-like has seven genes ζ2, Ψζ1, μ, Ψα, α2, α1, θ. ζ2 encodes for the ζ chain found in embryonic haemoglobins , α2, α1 encode for alpha chains, Ψζ1 and  Ψα are pseudogenes and the function of μ and θ is unknown. the β-like gene cluster consists of six genes ε, Gγ, Aγ, Ψβ, δ and β. ε gene encodes for the ε chain of haemoglobin Gower I and Gower II, Gγ and Aγ encode for γ chains of HbF, Ψβ is a pseudogene and δ and β encode fro chains found in HbA2 and HbA respectively (α2δ2 is HbA2 and not Hb Gower I that is erroneously shown in the diagram).

The earliest evidence of erythropoiesis is found in the extra-embryonic yolk sac in the form of blood islands. The erythrocytes produced by the yolk sac are larger, nucleated and contain embryonic haemoglobins (see N Engl J Med 1999; 340-617 for images of embryonic erythropoiesis). The yolk sac is the site for haematopoiesis from 19 days through week 8 of gestation.

The embryonic haemoglobin show co-operative oxygen binding but to a lesser extent than the adult haemoglobin. This results in a high affinity which reflect in a lower P50. The P50 of haemoglobin Gower I is 4mm, Gower II is 12mm and Portland is 6mm. The values for foetal haemoglobin is 19mm and adult haemoglobin is 26mm.

Deletions of ζ and ε gene causing thalassaemia have been described. These deletions include the α (in case of ζ gene) or β (in case of ε gene). Symptoms associated with these deletions are attributable to deletions of either α or β genes. There are no symptoms attributable to the ζ and ε genes.

Detection of ζ chains in the blood has a role in diagnosis of α thalassaemia.An ELISA using monoclonal antibodies against the ζ-globin chain for detection of southeast asian type of Hb-Bart’s hydrops foetalis has been described (Am J Clin Pathol 2008;129:309-315)



Development of erythrocytes involves coordinated changes in the nucleus and the cytoplasm of erythroid processors (see Morphology of Erythroid Precursors) . The proerythroblast is a large cell with a fine chromatin, the earliest forms not being different from other blasts. As the cell matures the chromatin becomes more clumped and the nucleus reduces in size and the cytoplasm becomes acidophilic. Finally a dark pyknotic nucleus is extruded from the orthchromatophilic normoblast to give a reticulocyte.

Figure 1. Basophilic Normoblast

A group of basophilic normoblast are shown above. The cytoplasm is basophilic and the chromatin more clumped than a proerythroblast. The two cells on the right are less mature than the two on the left.

Figure 2. A group of orthochromatophilic normoblasts

The figure above show a group of  orthochromatophilic normoblasts. The cells of the left are more mature. The nucleus is reduced to a dense body in the more mature forms. The cytoplasm still has a blue tinge. which contrasts from the megaloblasts shown below.

Figure 3. Basophilic, polychromatophilic and orthochromatophilic normoblasts

The figure above shows three stages of erythroid maturation. The cell on the top right is a basophilic normoblast, bottom right is a polychromatophilic normoblast and the bottom left is an orthochromatophlic normoblast. Note the evolution of nuclear and cytoplasmic changes.

Figure 4. Megaloblasts

The figure above shows a group of megaloblasts. Cells in the right lower corner have a cytoplasm which is fully haemoglobinized and resembles the mature erythrocyte. These cell still have a nucleus. The cell on the left upper corner has a cytoplasm resembling a orthochromatophilc normoblast (see figure 2 and 3) but nuclear features resembling a basophilic normoblast (figures 1 and 2). Megaloblastic anaemia results from conditions that hamper DNA synthesis (B12 deficiency, folate deficiency, Chemotherapy drugs). The nucleus of a megaloblast thus matures slower than the cytoplasm resulting in cells having a nuclear morphology resembling a previous stage. This, known as nucleo-cytoplasmic dissociation, is the characteristic feature of megaloblastic anaemia.