There are many naturally occurring, genetically determined variants of human haemoglobin (>1000)¹ and although many are harmless, some have serious clinical effects. Collectively, the clinical syndromes resulting from disorders of haemoglobin synthesis are referred to as ‘haemoglobinopathies’. They can be grouped into three main categories:

1. Those owing to structural variants of haemoglobin, such as Hb S.

2. Those owing to failure to synthesize one or more of the globin chains of haemoglobin at a normal rate, as in the thalassaemias.

3. Those owing to failure to complete the normal neonatal switch from fetal haemoglobin (Hb F) to adult haemoglobin (Hb A). The third category comprises a group of disorders referred to as hereditary persistence of fetal haemoglobin (HPFH).

An individual can also have a combination of two or more of these abnormalities.

Structural variants of haemoglobin

Alterations in the structure of haemoglobin are usually brought about by point mutations affecting one or, in some cases, two or more bases, coding for amino acids of the globin chains. An example of such a point mutation is Hb S caused by the substitution of valine for glutamic acid in position 6 of the β-globin chain (β6^Glu→Val). Less commonly, structural change is caused by shortening or lengthening of the globin chain. For example, five amino acids are deleted in the β chain of Hb Gun Hill, whereas in Hb Constant Spring 31 amino acids are added to the α chain. Mutations associated with a frame shift can also lead to synthesis of a structurally abnormal haemoglobin, which may be either shorter or longer than normal. There may also be combinations of segments of β and γ or δ chains resulting in hybrid haemoglobins; the β and δ combinations are known as the Lepore (5′δ3′β) and anti-Lepore (5′β3′δ) haemoglobins.

Many variant haemoglobins are haematologically and clinically silent because the underlying mutation causes no alteration in the function, solubility or stability of the haemoglobin molecule. Many of these variants are separated using electrophoresis or chromatography, but some are not and remain undetected. Some structural variants are associated with severe clinical phenotypes in the homozygous or even heterozygous state; these mutations affect the physical or chemical properties of the haemoglobin molecule, resulting in changes in haemoglobin solubility, stability or oxygen-binding properties. Some of these variants separate on electrophoresis or chromatography, whereas others do not. It is fortunate that the common haemoglobin variants that have clinical or genetic significance (e.g. Hbs S, C, D^Punjab, E and O^Arab) are readily detectable by electrophoretic and chromatographic techniques.

Haemoglobins with Reduced Solubility

Hb S

By far the most common haemoglobin variant in this group is sickle haemoglobin or Hb S. As a result of the replacement of glutamic acid by valine in position 6 of the β chain, Hb S has poor solubility in the deoxygenated state and can polymerize within the red cells. The red cell shows a characteristic shape change because of polymer formation and becomes distorted and rigid, the so-called sickle cell (see p. 90, Fig. 5.71). In addition, intracellular polymers lead to red cell membrane changes, generation of oxidant substances and abnormal adherence of red cells to vascular endothelium.

Clinical syndromes associated with common structural variants and those owing to their interaction with β thalassaemia are shown in Table 14.1.

Table 14.1 Clinical syndromes encountered with β^S and β^C variants

Sickle Cell Disease

Sickle cell disease ² (also called disorder) is a collective name for a group of conditions causing clinical symptoms that result from the formation of sickle red cells. It is common in people originating from Africa, but it is also found in considerable numbers of people of Indian, Arabic and Greek descent.

The homozygous state or sickle cell anaemia (β genotype SS) causes moderate anaemia resulting both from haemolysis and from the reduced oxygen affinity of haemoglobin S. The main clinical disability arises from repeated episodes of vascular occlusion by sickled red cells resulting in acute crises and eventually in end-organ damage. The clinical severity of sickle cell anaemia is extremely variable. This is partly due to the effects of inherited modifying factors, such as interaction with α thalassaemia or increased synthesis of Hb F and partly to socioeconomic conditions and other factors that influence general health.²

Sickle cell trait (β genotype AS), the heterozygous state, is very common, affecting millions of people worldwide. There are no associated haematological abnormalities. In vivo sickling occurs only at very high altitudes and at low oxygen pressures. Spontaneous haematuria, owing to sickling in the renal papillae, is found in about 1% of people with sickle cell trait.

Unstable Haemoglobins

Amino acid substitutions close to the haem group, or at the points of contact between globin chains, can affect protein stability and result in intracellular precipitation of globin chains. The precipitated globin chains attach to the red cell membrane giving rise to Heinz bodies, and the associated clinical syndromes were originally called the congenital Heinz body haemolytic anaemias. Changes in membrane properties may lead to haemolysis, often aggravated by oxidant drugs. There is considerable heterogeneity in the haematological and clinical effects of unstable haemoglobins. Many are almost silent and are detected only by specific tests, whereas others are severe, causing haemolytic anaemia in the heterozygous state. Hb Köln is the most common variant in this rare group of disorders.^3,⁴

Haemoglobins with Altered Oxygen Affinity

Haemoglobin variants with altered oxygen affinity are a rare group of variants that result in increased or reduced oxygen affinity.⁵ Mutations that increase oxygen affinity are generally associated with benign lifelong erythrocytosis. This may be confused with polycythaemia vera and may be inappropriately treated with cytotoxic drugs or ³²P.

Haemoglobin variants with decreased oxygen affinity are, with the exception of Hb S, even less common and are usually associated with mild anaemia and cyanosis. However, owing to the reduced oxygen affinity, these patients are not functionally anaemic despite the reduced Hb. The low steady-state haemoglobin concentration in Hb S homozygotes is, at least in part, a result of its reduced oxygen affinity.

Measurement of oxygen dissociation is described on p. 269.

Hb M

The Hb M group is another rare group of variants.⁶ Such haemoglobins have a propensity to form methaemoglobin, generated by the oxidation of ferrous iron in haem to ferric iron, which is incapable of binding oxygen. Despite marked cyanosis, there are few clinical problems. Most are associated with substitutions that disrupt the normal six-ligand state of haem iron.

Methaemoglobinaemia is also found in congenital NADH methaemoglobin reductase deficiency, as well as after exposure to oxidant drugs and chemicals (nitrates, nitrites, quinones, chlorates, phenacetin, dapsone and many others).

Thalassaemia syndromes

The thalassaemia syndromes ⁷ are a heterogeneous group of inherited conditions characterized by defects in the synthesis of one or more of the globin chains that form the haemoglobin tetramer. The clinical syndromes associated with thalassaemia arise from the combined consequences of inadequate haemoglobin production and of unbalanced accumulation of one type of globin chain. The former causes anaemia with hypochromia and microcytosis; the latter leads to ineffective erythropoiesis and haemolysis. Clinical manifestations range from completely asymptomatic microcytosis to profound anaemia that is incompatible with life and can cause death in utero (Table 14.2). This clinical heterogeneity arises as a result of the variable severity of the primary genetic defect in haemoglobin synthesis and the coinheritance of modulating factors, such as the capacity to synthesize increased amounts of Hb F.

Table 14.2 Clinical syndromes of thalassaemia

Clinically asymptomatic

Silent carriers

α⁺ thalassaemia heterozygotes (some cases)

Rare forms of β thalassaemia trait

Thalassaemia minor (low MCH and MCV, with or without mild anaemia)

α⁺ thalassaemia heterozygotes (some cases)

α° thalassaemia heterozygotes

α⁺ thalassaemia homozygotes

β° thalassaemia trait

β⁺ thalassaemia trait

Some cases of Hb E/β thalassaemia

Thalassaemia intermedia (transfusion independent)^a

Some β⁺/β⁺ thalassaemia homozygotes and compound heterozygotes

Interaction of β°/β°, β°/β⁺ or β⁺/β⁺ with

α thalassaemia

Interaction of β°/β or β⁺/β with triple α

Hb H disease

α°/Hb Constant Spring thalassaemia

β°/δβ or β⁺/δβ thalassaemia compound heterozygotes

δβ/δβ thalassaemia

Some cases of Hb E/β thalassaemia and Hb Lepore/β thalassaemia

Rare cases of heterozygotes for β thalassaemia mutation, particularly involving exon 3 (‘dominant β thalassaemia’)

Thalassaemia major (transfusion dependent)

β°/β° thalassaemia

β⁺/β⁺ thalassaemia

β°/β⁺ thalassaemia

Some cases of β°/Hb Lepore and β⁺/Hb Lepore thalassaemia

Some cases of β°/Hb E and β⁺/Hb E thalassaemia

MCH, mean cell haemoglobin; MCV, mean cell volume.

a β thalassaemia intermedia is defined as a symptomatic condition in which regular transfusion is not essential to maintain life; however some patients have a poor quality of life and may benefit from transfusion as their condition progresses.

Thalassaemias are generally inherited as alleles of one or more of the globin genes located on either chromosome 11 (for β, γ and δ chains) or on chromosome 16 (for α chains). They are encountered in every population in the world but are most common in the Mediterranean littoral and near equatorial regions of Africa and Asia. Gene frequencies for the α and β thalassaemias on a global basis range from 1% to more than 80% in areas where malaria is endemic.⁸

β Thalassaemia Syndromes

Many different mutations cause β thalassaemia and related disorders.⁹ These mutations can affect every step in the pathway of globin gene expression: transcription, processing of the messenger ribonucleic acid (mRNA) precursor, translation of mature mRNA and preservation of post-translational integrity of the β chain. More than 200 mutations have been described.¹⁰ Most types of β thalassaemia are the result of point mutations affecting the globin gene, but some large deletions are also known. Certain mutations are particularly common in some communities. This helps to simplify prenatal diagnosis, which is carried out by detection or exclusion of a particular mutation in fetal DNA.

The effect of different mutations varies greatly. At one end of the spectrum are a group of rare mutations, mainly involving exon 3 of the β globin gene, which are so severe that they can produce the clinical syndrome of thalassaemia intermedia in the heterozygous state. At the other end are mild alleles that produce thalassaemia intermedia in the homozygous or compound heterozygous state and some that are so mild that they are completely haematologically silent, with normal mean cell volume (MCV) and Hb A₂ in the heterozygous state. In between are the great majority of β⁺ and β° alleles, which cause β thalassaemia major in the homozygous or compound heterozygous state and in the heterozygous state give rise to a mild anaemia (or Hb at the low end of the normal range), with microcytic, hypochromic indices and raised Hb A₂.¹¹

β Thalassaemia major is a severe, transfusion-dependent, inherited anaemia. There is a profound defect of β chain production. Excess α chains accumulate and precipitate in the red cell precursors in the bone marrow resulting in ineffective erythropoiesis. The few cells that leave the marrow are laden with precipitated α chains and are rapidly removed by the reticuloendothelial system. The constant erythropoietic drive causes massive expansion of bone marrow and extramedullary erythropoiesis. If untreated, 80% of children with β thalassaemia die within the first 5 years of life.

Heterozygotes for β thalassaemia alleles usually have either a normal haemoglobin with microcytosis or a mild microcytic hypochromic anaemia; Hb A₂ is elevated and Hb F is sometimes also elevated. Laboratory features of various β thalassaemia syndromes are shown in Table 14.3.

Table 14.3 Laboratory findings in thalassaemia

α Thalassaemia Syndromes

There are four syndromes of α thalassaemia:¹² α⁺ thalassaemia trait, where one of the two globin genes on a single chromosome fails to function; α° thalassaemia trait, where two genes on a single chromosome fail to function; Hb H disease, with three genes affected; and Hb Bart’s hydrops fetalis, where all four are absent or defective. These syndromes are usually a result of deletions of one or more genes, although approximately 20% of the mutations described are non-deletional. α⁺ thalassaemia is particularly common in Africa and α° thalassaemia is common in South-east Asia. The laboratory features are shown in Table 14.3.

Hb Bart’s hydrops fetalis occurs mainly in people from South-east Asia but is also occasionally observed in people from Greece, Turkey and Cyprus. An affected fetus will be stillborn or will die shortly after birth. Severe anaemia and oedema are the hallmarks of this condition. Women carrying a hydropic fetus have a high incidence of complications of pregnancy. Prenatal diagnosis should be offered for women at risk of having a fetus with Hb Bart’s hydrops fetalis.

Hb H disease gives rise to haemolytic anaemia; patients rarely require transfusion or splenectomy.

α° thalassaemia trait is characterized by microcytic, hypochromic indices. The haemoglobin level may be normal or slightly reduced. α⁺ thalassaemia trait can be completely silent or there may be borderline microcytosis with a slightly reduced or normal mean cell haemoglobin concentration (MCHC). Haematologically, homozygosity for α⁺ thalassaemia trait resembles heterozygosity for α° thalassaemia trait, but the genetic implications are very different. Both α⁺ thalassaemia trait and α° thalassaemia trait are more difficult to diagnose than β thalassaemia trait because there is no characteristic elevation in Hb A₂, and Hb H bodies are frequently not demonstrated. Definitive diagnosis of the α thalassaemia trait is more reliably made with the use of DNA techniques or globin chain biosynthesis studies.

Thalassaemic Structural Variants

These are abnormal haemoglobins characterized by both a biosynthetic defect and an abnormal structure, such as the Lepore haemoglobins (Table 14.4).

Table 14.4 Thalassaemic structural variants^a

Increased Hb F in Adult Life

Haemoglobin production in man is characterized by two major switches in the haemoglobin composition of the red cells. During the first 3 months of gestation, human red cells contain embryonic haemoglobins (see p. 302), whereas during the last 6 months of gestation, red cells contain predominantly fetal haemoglobin. The major transition from fetal to adult haemoglobin synthesis occurs in the perinatal period and by the end of the first year of life red cells have a haemoglobin composition that usually remains constant throughout adult life. The major haemoglobin is then Hb A, but there are small amounts of Hb A₂ and Hb F. Only 0.2–1.0% of total haemoglobin in human red cells is Hb F and it is restricted to a few cells called ‘F’ cells. Both the number of F cells and the amount of Hb F per cell can be increased in various conditions, particularly if there is rapid bone marrow regeneration.¹³

The general organization of human globin gene clusters is shown in Figure 14.1. The products of two γ genes differ in only one amino acid: ^Gγ has glycine in position 136, whereas ^Aγ has alanine. In fetal red cells, the ratio of ^Gγ to ^Aγ is approximately 3:1; in adult red cells, it is approximately 2:3.

In recent years there has been much interest in the attempts to manipulate the fetal switch pharmacologically. If it was possible to reactivate Hb F synthesis reliably beyond the perinatal period, both β thalassaemia major and sickle cell disease would be ameliorated.

Inherited Abnormalities That Increase Hb F Concentration

More than 50 mutations that increase Hb F synthesis have been described.^13,¹⁴ They result in one of two phenotypes, HPFH or δβ thalassaemia; differentiation between these two types is not always simple but has clinical relevance. In general, HPFH has a higher percentage of Hb F and much more balanced chain synthesis. The most common, the African type of HPFH, is associated with a high concentration of Hb F (15–45%), pancellular distribution of Hb F on Kleihauer staining and normal red cell indices. Mutations causing increased synthesis of Hb F are mostly deletions, but some non-deletion mutations have also been described. In contrast, subjects with δβ thalassaemia have lower levels of Hb F accompanied by microcytic, hypochromic indices. The major clinical significance of these abnormalities is their interaction with β thalassaemia and Hb S. Compound heterozygotes for either of these conditions and HPFH have much milder clinical syndromes than the homozygotes for haemoglobin S or β thalassaemia. Compound heterozygotes for either of these conditions and δβ thalassaemia have a condition much closer in severity to the homozygous states.

Increased Hb F is also found in many other haematological conditions, including congenital red cell aplasia and congenital aplastic anaemia (Blackfan-Diamond and Fanconi anaemia, respectively), juvenile myelomonocytic leukaemia and some myelodysplastic syndromes. A small but significant increase in Hb F may occur in the presence of erythropoietic stress (haemolysis, bleeding, recovery from acute bone marrow failure) and in pregnancy.

Investigation of patients with a suspected haemoglobinopathy

Investigation of a person at risk of a haemoglobinopathy encompasses the confirmation or exclusion of the presence of a structural variant, thalassaemia trait or both. If a structural haemoglobin variant is present, it is necessary to ascertain the clinical significance of the particular variant so that the patient is appropriately managed. If it is confirmed that thalassaemia trait is present, it is not usually necessary to determine the precise mutation present because the clinical significance is usually negligible. The exception to this is an antenatal patient whose partner has also been found to have thalassaemia trait. If prenatal diagnosis is being considered, it may be necessary to undertake mutation analysis to predict fetal risk accurately and to facilitate prenatal diagnosis (see p. 146).

Because the inheritance of a haemoglobinopathy per se has genetic implications, it is important that genetic counselling is available for these patients.

In the majority of patients, the presence of a haemoglobinopathy can be diagnosed with sufficient accuracy for clinical purposes from knowledge of the patient’s ethnic origin and clinical history (including family history) and the results of physical examination combined with relatively simple haematological tests. Initial investigations should include determination of haemoglobin concentration and red cell indices. A detailed examination of a well-stained blood film should be carried out. In some instances, a reticulocyte count and a search for red cell inclusions give valuable information. Assessment of iron status by estimation of serum iron and total iron binding capacity and/or serum ferritin is sometimes necessary to exclude iron deficiency. Other important basic tests are haemoglobin electrophoresis or high-performance liquid chromatography (HPLC), a sickle solubility test and measurement of Hb A₂ and Hb F percentage. In cases of common haemoglobin variants and classical β thalassaemia trait, accurate data from these tests will facilitate a reliable diagnosis without the need for more sophisticated investigations. However, definitive diagnosis of some thalassaemia syndromes can only be obtained using DNA technology (see p. 147 and p. 330). Similarly, in particular situations, haemoglobin variants will require unequivocal identification by the use of DNA technology or protein analysis by mass spectrometry.¹⁵ Individuals or families who require such investigation must be carefully selected on the basis of family history and on the results of the basic investigations described later in this chapter. Large-scale screening programmes are increasingly being undertaken in some countries where individual case histories and the results of other laboratory tests are not usually available. The problems of such programmes are discussed on p. 318.

The majority of errors occurring in the detection and identification of a haemoglobinopathy are the result of either failure to obtain correct laboratory data or failure to interpret data correctly. In this chapter, a sequence of investigations is proposed based on procedures that should be available in the laboratory of any major hospital. Automated HPLC is increasingly replacing haemoglobin electrophoresis as the initial investigative procedure in laboratories analysing large numbers of samples. Isoelectric focusing (IEF) is, in general, used only to a limited extent, mainly for neonatal screening or in specialist laboratories and it is only briefly described here.

Laboratory investigation of a suspected haemoglobinopathy should follow a defined protocol, which should be devised to suit individual local requirements. The data obtained from the clinical findings, blood picture and electrophoresis or HPLC will usually indicate in which direction to proceed. The investigation for a structural variant is described in the first section and that for a suspected thalassaemia syndrome is described in the second section of this chapter. Screening tests for thalassemia trait and haemoglobin E trait that may be especially applicable in under-resourced areas are described in Chapter 26.

Laboratory detection of haemoglobin variants

A proposed scheme of investigation is shown in Figure 14.2 and a list of procedures follows:^15,¹⁶

1. Blood count and film examination (see below)

2. Collection of blood and preparation of haemolysates (see p. 309)

3. Cellulose acetate electrophoresis, Tris buffer, pH 8.5 (see p. 310)

4. Citrate agar or acid agarose gel electrophoresis, pH 6.0 (see p. 312)

5. Automated HPLC (see pp. 314)

6. IEF (see p. 315)

7. Tests for Hb S (see p. 315)

8. Detection of unstable haemoglobins (see p. 318)

9. Detection of Hb Ms (see p. 319)

10. Detection of altered affinity haemoglobins (see p. 320)

11. Differentiation of common structural variants (see p. 320)

12. Neonatal screening (see p. 318)

13. Tests, such as zinc protoporphyrin estimation, to exclude iron deficiency as a cause of microcytosis (see Chapter 9)

14. Molecular techniques (see Chapter 8)

15. Procedures for use in under-resourced laboratories (see Chapter 26).

Figure 14.2 Suggested scheme of investigation for structural variants.

Globin chain electrophoresis, pH 8.0 and 6.3, is now rarely performed. Methods can be found in previous editions of this book.

Blood Count and Film

The blood count, including Hb and red cell indices, provides valuable information useful in the diagnosis of both α and β thalassaemia interactions with structural variants (see Chapter 3). A film examination may reveal characteristic red cell changes such as target cells in Hb C trait, sickle cells in sickle cell disease and irregularly contracted cells in the presence of an unstable haemoglobin (see Chapter 5).

Discriminant functions using various formulae have been proposed as a basis for further testing for thalassaemia,^17,¹⁸ but we do not advise their use. Although such functions and formulae do indicate whether thalassaemia or iron deficiency is more likely, they may lead to individuals who have both iron deficiency and thalassaemia trait not being tested promptly. Generally this is not a problem and indeed it may be preferable to keep the patient under observation until iron deficiency has been treated and then to reassess the likelihood of thalassaemia trait. However, many of the patients who require testing for thalassaemia are women who are already pregnant. In such patients the likely delay in testing is unacceptable. Moreover, these formulae do not appear to have been validated for use during pregnancy. For these reasons, we advise that whenever genetic counselling might be required, testing for β thalassaemia trait should be carried out in all individuals with an MCH <27 pg and screening for α thalassaemia trait should be carried out in those individuals with an MCH <25 pg who belong to an ethnic group in which α° thalassaemia is prevalent.¹¹

Collection of Blood and Preparation of Haemolysates

EDTA is the most convenient anticoagulant because it is used for the initial full blood count and film (see Chapter 1), although samples taken into any anticoagulant are satisfactory. Cells freed from clotted blood can also be used.

Preparation of Haemolysate for Qualitative Haemoglobin Electrophoresis

See individual methods.

Preparation of Haemolysate for the Quantification of Haemoglobins and Stability Tests

Preparation of haemolysate for the quantification of haemoglobins and stability tests can be used for qualitative electrophoresis and is necessary for quantitation of Hb A₂ and F or variant haemoglobins by elution. It is also essential for reliable stability tests.

Lyse 2 volumes of washed packed cells in 1 volume of distilled water and then add 1 volume of carbon tetrachloride (CCl₄). Alternatively, lyse by freezing and thawing, then add 2 volumes of CCl₄. Shake the tubes vigorously for approximately 1 min, then centrifuge at 1200 g (3000 rev/min) for 30 min at 4°C. Transfer the supernatant to a clean sample container and adjust the Hb to 100 ± 10 g/l with water. If an unstable Hb is suspected organic solvents should be avoided.

Note: Whole blood samples are best stored as washed, packed cells frozen as droplets in liquid nitrogen and subsequently stored at −20°C, −70°C or over liquid nitrogen. Alternatively, haemolysates may also be frozen at −20°C, −70°C or over liquid nitrogen.

Control Samples

Interpretation of migration patterns of test samples is undertaken by comparison to migration and separation of known abnormal haemoglobins used as control materials. Ideally, a mixture of Hbs A, F, S and C should be included on each electrophoretic separation. This material can be prepared as follows:

1. The control can be made from either the combination of a sickle cell trait sample (Hb A + Hb S) combined with a Hb C trait sample (Hb A + Hb C) and normal cord blood (Hb F + Hb A) or the combination of normal cord blood with a sample from a person with Hb SC disease (Hb S + Hb C).

2. Prepare lysates by the method given for a purified haemolysate.

3. Mix equal volumes of the lysates together and add a few drops of 0.3 mol/l KCN (20 g/l).

4. Analyse samples by electrophoresis to assess quality.

5. Aliquot and store frozen.

Note: Repeated freezing and thawing should be avoided.

Lyophilized controls are stable for considerably longer than liquid and can be purchased from commercial sources.

Quality Assurance

Because the haemoglobinopathies are inherited conditions, some of which carry considerable clinical and genetic implications, precise documentation and record-keeping are of paramount importance.¹⁹ The use of cumulative records when reviewing a patient’s data is very useful because it of itself constitutes an aspect of quality assurance. In some situations, repeat sampling, family studies or both may be required to elucidate the nature of the abnormality in an individual.

In-house standard operating procedures should be followed carefully, particularly in this field of haematology, where a small difference in technique can make a significant difference in the results obtained and can lead to misdiagnosis. Many of the techniques described have attention drawn to specific technical details that are important for ensuring valid results.²⁰

It is necessary to use reference standards and control materials in each of the analyses undertaken and in some cases to use duplicate analysis to demonstrate precision. There are international standards for Hb F and Hb A₂ (see p. 591), whereas in some countries national reference preparations are also available from national standards institutions. These are extremely valuable because the target values have been established by collaborative studies. Control materials can be prepared in-house or obtained commercially. Samples stored as whole blood at 4°C can be used reliably for several weeks. All laboratories should confirm the normal range for their particular methods and the normal range obtained should not differ significantly from published data.

All laboratories undertaking haemoglobin analysis should participate in an appropriate proficiency testing programme (see p. 594). In the UK, the National External Quality Assessment Scheme (NEQAS) provides samples for sickle solubility tests; for detection and quantitation of variant haemoglobins; and for quantitation of Hbs A₂, F and S.

National and international guidelines have been published for all aspects of the investigations given here.^11,^21–23,^20,^24,²⁵

Cellulose Acetate Electrophoresis at Alkaline pH

Haemoglobin electrophoresis at pH 8.4–8.6 using cellulose acetate membrane is simple, reliable and rapid. It is satisfactory for the detection of most common clinically important haemoglobin variants.²³^–²⁵

Principle

At alkaline pH, haemoglobin is a negatively charged protein and when subjected to electrophoresis will migrate toward the anode (+). Structural variants that have a change in the charge on the surface of the molecule at alkaline pH will separate from Hb A. Haemoglobin variants that have an amino acid substitution that is internally sited may not separate and those that have an amino acid substitution that has no effect on overall charge will not separate by electrophoresis.

Equipment

Electrophoresis tank and power pack. Any horizontal electrophoresis tank that will allow a bridge gap of 7 cm. A direct current power supply capable of delivering 350 V at 50 mA is suitable for both cellulose acetate and citrate agar electrophoresis

Wicks of filter or chromatography paper

Blotting paper

Applicators. These are available from most manufacturers of electrophoresis equipment, but fine microcapillaries are also satisfactory

Cellulose acetate membranes. Plastic-backed membranes (7.6 × 6.0 cm) are recommended for ease of use and storage

Staining equipment.

Reagents

Electrophoresis buffer. Tris/EDTA/borate (TEB), pH 8.5. Tris-(hydroxymethyl)aminomethane (Tris), 10.2 g; EDTA (disodium salt), 0.6 g; boric acid, 3.2 g; water to 1 litre. The buffer should be stored at 4°C and can be used up to 10 times without deterioration

Wetting agent. For example, Zip-prep solution (Helena Laboratories): 1 drop of Zip-prep in 100 ml water

Fixative/stain solution. Ponceau S, 5 g; trichloroacetic acid, 7.5 g; water to 1 litre

Destaining solution. 3% (v/v) acetic acid, 30 ml; water to 1 litre

Haemolysing reagent. 0.5% (v/v) Triton X-100 in 100 mg/l potassium cyanide.

Method

1. Centrifuge samples at 1200 g for 5 min. Dilute 20 μl of the packed red cells with 150 μl of the haemolysing reagent. Mix gently and leave for at least 5 min. If purified haemolysates are used, dilute 40 μl of 10 g/dl haemolysate with 150 μl of lysing reagent.

2. With the power supply disconnected, prepare the electrophoresis tank by placing equal amounts of TEB buffer in each of the outer buffer compartments. Wet two chamber wicks in the buffer and place one along each divider/bridge support ensuring that they make good contact with the buffer.

3. Soak the cellulose acetate by lowering it slowly into a reservoir of buffer. Leave the cellulose acetate to soak for at least 5 min before use.

4. Fill the sample well plate with 5 μl of each diluted sample or control and cover with a 50 mm coverslip or a ‘short’ glass slide to prevent evaporation. Load a second sample well plate with Zip-prep solution.

5. Clean the applicator tips immediately prior to use by loading with Zip-prep solution and then applying them to a blotter.

6. Remove the cellulose acetate strip from the buffer and blot twice between two layers of clean blotting paper. Do not allow the cellulose acetate to dry.

7. Load the applicator by depressing the tips into the sample wells twice and apply this first loading onto some clean blotting paper. Reload the applicator and apply the samples to the cellulose acetate.

8. Place the cellulose acetate plates across the bridges, with the plastic side uppermost. Place two glass slides across the strip to maintain good contact. Electrophorese at 350 V for 25 min.

9. After 25 min electrophoresis, immediately transfer the cellulose acetate to Ponceau S and fix and stain for 5 min.

10. Remove excess stain by washing for 5 min in the first acetic acid reservoir and for 10 min in each of the remaining two. Blot once, using clean blotting paper and leave to dry.

11. Label the membranes and store in a protective plastic envelope.

Interpretation and Comments

Figure 14.3 shows the relative electrophoretic mobilities of some common haemoglobin variants at pH 8.5 on cellulose acetate. Satisfactory separation of Hbs C, S, F, A and J is obtained (Fig. 14.4). In general, Hbs S, D and G migrate closely together, as do Hbs C, E and O^Arab. Differentiation between these haemoglobins can be obtained by using acid agarose gels, citrate agar electrophoresis, HPLC or IEF. However, there are slight differences in mobility between Hbs S, Lepore and D^Punjab and also between Hbs C and E; optimization of the technique will facilitate detection of the difference. Generally, the Lepore Hbs and Hb D^Punjab migrate slightly anodal to Hb S (i.e. they are slightly faster than S); Hb C migrates slightly cathodal to Hb E (i.e. it is slightly slower than E).

Figure 14.3 Schematic representation of relative mobilities of some abnormal haemoglobins. Cellulose acetate, pH 8.5.

Figure 14.4 Relative mobilities of some abnormal haemoglobins. Cellulose acetate, pH 8.5.

All samples showing a single band in either the S or C position should be analysed further using acid agarose or citrate agar gel electrophoresis, HPLC or IEF to exclude the possibility of a compound heterozygote such as SD, SG, CE or CO^Arab.

The quality of separation resulting from this procedure is affected primarily by both the amount of haemoglobin applied and the positioning of the origin. Also, delays between application of the sample and commencement of the electrophoresis, delay in staining after electrophoresis or inadequate blotting of the acetate prior to application will cause poor results. This technique is sensitive enough to separate Hb F from Hb A and to detect Hb A₂ variants.

If an abnormal haemoglobin is present, the detection of a Hb A₂ variant band in conjunction with the abnormal fraction is evidence that the variant is an α chain variant. Globin electrophoresis at both acid and alkaline pH is also useful in elucidating which globin chain is affected. However, with the more ready availability of HPLC, it is less often needed.

When an abnormal haemoglobin is found, it may be of diagnostic importance to measure the percentage of the variant; this can be done by the electrophoresis with elution procedure for Hb A₂ estimation given on p. 322. Quantitation of Hb S is often clinically useful, both in patients with sickle cell disease who are being treated by transfusion and for the diagnosis of conditions in which Hb S is coinherited with α and β thalassaemia, as outlined in Table 14.5. Quantitation of Hb S can be done with HPLC, electrophoresis with elution or by microcolumn chromatography.

Table 14.5 Results of laboratory investigations in interactions of Hb S and α or β thalassaemia in adults

Citrate Agar Electrophoresis at pH 6.0

Equipment ^26,²⁷

Cooling bars (Helena Laboratories)

Perspex trays, 80 × 100 × 2 mm

Wicks of sponge, filter paper or chromatography paper

Applicators. These are available from most manufacturers of electrophoresis equipment, but fine microcapillaries are also satisfactory.

Reagents

Difco Bacto-agar.

Lysing reagent. 0.5% (v/v) Triton X-100 in 100 mg/l potassium cyanide.

Buffer

Stock buffer. Trisodium citrate dihydrate, 73.5 g; 0.5 mol/l citric acid, 34.0 ml; water to 1 litre

Working buffer. Dilute stock buffer 1 in 5 with water. Prepare on day of use

5 g/dl potassium cyanide. Potassium cyanide, 0.5 g; distilled water to 100 ml

Dianisidine stain. 3% hydrogen peroxide (10 volumes), 1.0 ml; 1% sodium nitroprusside (nitroferricyanide), 1.0 ml; 3% acetic acid, 10.0 ml; 0.2% o-dianisidine in methanol, 5.0 ml. Prepare mixture just before use

Wetting agent. For example, Zip-prep solution (Helena Laboratories); 1 drop in 100 ml water.

3% Acetic acid. 120 ml glacial acetic acid made up to 4 litres with water

Gel-Bond or similar support.

Method

1. Centrifuge sample (1200 g for 5 min). Dilute 20 μl of packed red cells with 300 μl of haemolysing reagent. Mix gently and leave for at least 5 min. For cord blood samples, dilute 20 μl of packed red cells with 150 μl of the lysing reagent. For purified haemolysates, dilute 20 μl of 10 g/dl haemolysate with 150 μl of lysing reagent.

2. With the power supply disconnected, place equal volumes working buffer in each of the outer buffer compartments. Wet two sponge wicks in the buffer and place one in each compartment against the divider. Place two frozen cooling sticks in each central chamber. If cooling bars or ice packs are not available, run the electrophoresis at 4°C.

3. Add 0.5 g agar to 50 ml working buffer. Heat to approximately 95°C, stirring gently until the agar has dissolved. Allow to cool to 60°C; add 0.5 ml of 5 g/dl potassium cyanide. Pipette approximately 10 ml into each of four Perspex trays (80 × 100 × 2 mm) and allow to stand for about 15 min at room temperature until set. These gels may be kept for 1 week at 4°C in a sealed plastic bag. Allow gels to come to room temperature before use.

4. Fill the sample well plate with 5 μl of each sample and cover with a 6 cm coverslip or glass slide. Load a second sample well plate with Zip-prep solution. Clean the applicator tips by loading with Zip-prep solution and then applying them to blotting paper. Load the applicator and apply this first loading onto some clean blotting paper. Reload the applicator and apply the samples to the agar gel.

5. Place gel plate in an inverted position in the electrophoresis tank so that the gel is in contact with sponge wicks and run at a constant voltage of 50 V for 60 min.

6. After 60 min, disconnect from power supply, remove gel and apply the stain solution by layering onto the agar using a Pasteur pipette. Allow to stain for 10 min at room temperature.

7. Wash in three changes of 3% acetic acid, float gels onto the hydrophilic side of a piece of Gel Bond and leave to dry. These mounted gels may then be kept indefinitely.

Interpretation and Comments

Figure 14.5 shows the relative electrophoretic mobilities of some common haemoglobin variants at pH 6.0 on citrate agar.

Figure 14.5 Schematic representation of relative mobilities of some abnormal haemoglobins. Citrate agar pH 6.0.

Agarose Gel Electrophoresis

Agarose gels are commercially available as substitutes for both alkaline and acid separation systems. They are simple to use and particularly useful in laboratories that process small numbers of samples.

Reagents and Method

The manufacturer’s method should be followed.

Interpretation

With acid agarose systems, the principle of the test is the same as that of citrate agar electrophoresis at the same pH, but it should be noted that there are significant differences in mobility of some variant haemoglobins. With alkaline systems, in general the same separation patterns are obtained, but where individual application notes are available these should be used for reference. Because not all kits provide these, laboratories may need to build up their own data on known variants.

Automated High-Performance Liquid Chromatography

Automated cation-exchange HPLC ²⁸ is being used increasingly as the initial diagnostic method in haemoglobinopathy laboratories with a high workload.²⁹ Both capital and consumable costs are higher than with haemoglobin electrophoresis, but labour costs are less; overall costs may be similar.³⁰ In comparison with haemoglobin electrophoresis, HPLC has four advantages:

1. The analysers are automated and thus utilize less staff time and permit processing of large batches.

2. Very small samples (5 μl) are sufficient for analysis; this is especially useful in paediatric work.

3. Quantification of normal and variant haemoglobins is available on every sample.

4. A provisional identification of a larger proportion of variant haemoglobins can be made.³¹

Principle

HPLC depends on the interchange of charged groups on the ion exchange material with charged groups on the haemoglobin molecule. A typical column packing is 5 μm spherical silica gel. The surface of the support is modified by carboxyl groups to have a weakly cationic charge, which allows the separation of haemoglobin molecules with different charges by ion exchange. When a haemolysate containing a mixture of haemoglobins is adsorbed onto the resin, the rate of elution of different haemoglobins is determined by the pH and ionic strength of any buffer applied to the column. With automated systems now in use, elution of the charged molecules is achieved by a continually changing salt gradient; fractions are detected as they pass through an ultraviolet/visible light detector and are recorded on an integrating computer system. Analysis of the area under these absorption peaks gives the percentage of the fraction detected. The time of elution (retention time) of any normal or variant haemoglobin present is compared with that of known haemoglobins, providing quantification of both normal haemoglobins (A, F and A₂) and many variants.

Figure 14.6 shows a schematic representation of an HPLC system and Figure 14.7 shows a chromatogram of a mixture of different haemoglobins. Systems are available from various manufacturers.

Figure 14.6 Diagrammatic representation of high-performance liquid chromatography (HPLC) showing the flow of sample and buffers.

Figure 14.7 A mixture of haemoglobins separated by high-performance liquid chromatography (HPLC).

Method

The manufacturer’s procedure should be followed. To prolong the life of the column it is important to follow the manufacturer’s instructions with regard to the concentration of haemoglobin in the sample to be injected.

Interpretation and Comments

Results are accurate and reproducible, but as with every method of haemoglobin analysis, controls should be run with every batch. If the system is being used for the detection of haemoglobin variants, elution times can be compared with those of known controls; actual times, however, are affected by the batch of buffer and column, the age of the column and the laboratory temperature. A better comparison may be obtained using the relative elution time, which is calculated by dividing the elution time of the variant with that of the main Hb A fraction. It should be noted that Hb A is separated into its component fractions of A₀ and A₁ and the A₁ fraction frequently subdivides into several peaks. Skill is required in interpretation of the results because various normal and abnormal haemoglobins may have the same retention time and a glycosylated variant haemoglobin will have a different retention time from the non-glycosylated form. HPLC usually separates Hbs A, A₂, F, S, C, D^Punjab and G^Philadelphia from each other.^29,³² However, both Hb E and Hb Lepore co-elute with haemoglobin A₂ (as other haemoglobins co-elute with A, S and F). The retention time of glycosylated and other derivatives of Hb S can be the same as those of Hb A₀ and A₂. For example, derivatives of haemoglobin S co-elute with haemoglobin A₂, so that percentages of A₂ by this method are inaccurate and therefore do not have the same significance as percentage of haemoglobin A₂ measured by alternative methods.³³ For these reasons, and because there are more than 1000 variants identified, HPLC can never definitively identify any haemoglobin. It is important to analyse variants found using second-line techniques, such as a sickle solubility test, alkaline and acid electrophoresis, or iso-electric focusing.

HPLC is also applicable for the quantification of Hb A_1c for the monitoring of diabetes mellitus; to make optimal use of staff and equipment, this procedure is sometimes carried out in haematology laboratories. In fact, an increased glycosylated fraction is not infrequently noted when HPLC is performed for investigation of a suspected haemoglobinopathy.³⁴ Unless the patient is already known to suffer from diabetes mellitus, this abnormality should be drawn to the attention of clinical staff.

Isoelectric Focusing

Principle

IEF utilizes a matrix containing carrier ampholytes of low molecular weight and varying isoelectric points (pIs). These molecules migrate to their respective pIs when a current is applied, resulting in a pH gradient being formed; for haemoglobin analysis, a pH gradient of 6–8 is usually used. Haemoglobin molecules migrate through the gel until they reach the point at which their individual pIs equal the corresponding pH on the gel. At this point, the charge on the haemoglobin is neutral and migration ceases. The pH gradient counteracts diffusion and the haemoglobin variant forms a discrete narrow band.^35,³⁶

Method

Pre-prepared plates of either polyacrylamide or agarose gel can be obtained from various manufacturers. For the exact method, the manufacturer’s instructions should be followed.

Interpretation and Comments

IEF is satisfactory for analysis of haemolysates, whole blood samples or dried blood spots. The use of dried blood spots is suitable for samples that have to be transported long distances and where only a few drops of blood can be obtained. Whereas IEF has the advantage that it separates more variants than cellulose acetate, it also has the disadvantage that it separates haemoglobin into its post-translational derivatives. For instance, Hb F separates into F₁ (acetylated F) and F₁₁; Hb A can produce five bands – A₀, A₁, A(αmet), A(βmet) and A(αβmet) – and similarly for other haemoglobins. This makes interpretation more difficult. Identification of variants is still only provisional using IEF, and second-line methods should be used for further analysis.

Figure 14.8 shows the relative isoelectric points of some common haemoglobin variants and Figure 14.9 shows the separation obtained.

Figure 14.8 Schematic representation of relative mobilities of some abnormal haemoglobins. Isoelectric focusing. Note the presence of two bands for haemoglobin Bart’s. The scale in mm is not linear. *Indicates α chain mutations.

(Reproduced with permission from Basset et al.³⁶)

Figure 14.9 Relative mobilities of some abnormal haemoglobins. Isoelectric focusing. The controls are mixtures of various haemoglobins.

Tests for Hb S

Tests to detect the presence of Hb S depend on the decreased solubility of this haemoglobin at low oxygen tensions.

Sickling in Whole Blood

The sickling phenomenon may be demonstrated in a thin wet film of blood (sealed with a petroleum jelly/paraffin wax mixture or with nail varnish). If Hb S is present, the red cells lose their smooth, round shape and become sickled. This process may take up to 12 h in Hb S trait, whereas changes are apparent in homozygotes and compound heterozygotes after 1 h at 37°C. These changes can be hastened by the addition of a reducing agent such as sodium dithionite as follows:

Reagents

Disodium hydrogen phosphate (Na₂HPO₄). 0.114 mol/l (16.2 g/l)

Sodium dithionite (Na₂S₂O₄). 0.114 mol/l (19.85 g/l). Prepare freshly just before use

Working solution. Mix 3 volumes of Na₂HPO₄ with 2 volumes of Na₂S₂O₄ to obtain a pH of 6.8 in the resultant solution. Use immediately.

Method

Add 5 drops of the freshly prepared reagent to 1 drop of anticoagulated blood on a slide. Seal between slide and coverglass with a petroleum jelly/paraffin wax mixture or with nail varnish. Sickling takes place almost immediately in sickle cell anaemia and should be obvious in sickle cell trait within 1 h (Fig. 14.10). A test on a positive control of Hb A plus Hb S must be performed at the same time.

Figure 14.10 Photomicrograph of sickled red cells. Sickle cell anaemia. Sealed preparation of blood. Fully sickled filamentous forms predominate.

Hb S Solubility Test

Principle

Sickle cell haemoglobin is insoluble in the deoxygenated state in a high molarity phosphate buffer. The crystals that form refract light and cause the solution to be turbid.³⁷

Reagents

Phosphate buffer. Anhydrous dipotassium hydrogen phosphate, 215 g; anhydrous potassium dihydrogen phosphate, 169 g; sodium dithionite, 5 g; saponin, 1 g; water to 1 litre.

Note: Dissolve the K₂HPO₄ in water before adding the KH₂PO₄, then add the dithionite and finally the saponin. This solution is stable for 7 days. Store refrigerated.

Method

1. Pipette 2 ml of reagent into three 12 × 75 mm test tubes.

2. Allow the reagent to warm to room temperature.

3. Add 10 μl of packed cells (from EDTA-anticoagulated blood) to one tube, 10 μl of packed cells from a known sickle cell trait subject as a positive control to the second tube and 10 μl packed cells from a normal subject as a negative control to the final tube.

4. Mix well and leave to stand for 5 min.

Note: The blood reagent mixture should be light pink or red. A light orange colour indicates that the reagent has deteriorated.

5. Hold tube 2.5 cm in front of a white card with narrow black lines and read for turbidity, in comparison with the positive and negative control samples.

6. If the test appears to be positive, centrifuge at 1200 g for 5 min. A positive test will show a dark red band at the top, whereas the solution below will be pink or colourless.

Interpretation and Comments

A positive solubility or sickling test indicates the presence of Hb S and as such is useful in the differential diagnosis of Hbs D and G, which migrate with Hb S on cellulose acetate electrophoresis at alkaline pH. Positive results are also obtained on samples containing the rare haemoglobins that have both the Hb S mutation and an additional mutation in the β chain. A positive solubility test merely indicates the presence of a sickling haemoglobin and does not differentiate between homozygotes, compound heterozygotes and heterozygotes. In an emergency, it may be necessary to decide if an individual suffers from sickle cell disease before the haemoglobin electrophoresis results are available. In these circumstances, if the solubility test is positive, a provisional diagnosis of sickle cell trait can be made if the red cell morphology is normal on the blood film. If the blood film shows any sickle cells or numerous target cells, irrespective of the Hb, a provisional diagnosis of sickle cell disease should be made; many patients with sickle cell/Hb C compound heterozygosity will have a normal Hb. Remember that the sickle test is likely to be negative in infants with sickle cell disease.

False-positive results have been reported in severe leucocytosis; in hyperproteinaemia (such as multiple myeloma); and in the presence of an unstable haemoglobin, especially after splenectomy. The use of packed cells, as described in this method, minimizes the problem of false-positive results caused by hyperproteinaemia and hyperlipidaemia.

False-negative results can occur in patients with a low Hb and the use of packed cells will overcome this problem. False-negative results may also occur if old or outdated reagents are used and if the dithionite/buffer mixture is not freshly made. False-negative results are likely to be found in infants younger than age 6 months and in other situations (e.g. post-transfusion), in which the Hb S level is <20%.

All sickle tests, whether positive or negative, must be confirmed by electrophoresis or HPLC at the earliest opportunity.

Neonatal (newborn) screening

Cord blood or a heel prick sample should be tested from all babies at risk of sickle cell disease or β thalassaemia major (i.e. where the mother has a gene for Hb S, C, D^Punjab, E, O^Arab, Lepore or β or δβ thalassaemia trait). If a cord-blood specimen is used, it is important that the sample is collected by venepuncture of the cleaned umbilical vein to avoid contamination with maternal blood because even small quantities of maternal blood can cause a case of sickle cell disease to be misdiagnosed as sickle cell trait.

In areas where the frequency of haemoglobinopathies is high, universal neonatal (newborn) screening should be undertaken where possible. Universal neonatal screening is now being carried out in England and will be gradually extended to the rest of the UK.³⁸ The screening programme is linked to the existing dried blood spot screening programme in place for phenylketonuria and congenital hypothyroidism. The same dried blood spot sample is tested for sickle cell disease. It must be emphasized that the main function of this screening is to detect sickle cell disease, although many cases of β thalassaemia major are also detected, dependent on the mutations present. Dried blood spot samples are tested using HPLC and IEF – HPLC is typically the first-line test and abnormalities are confirmed by IEF. Haemoglobin electrophoresis is not recommended for the analysis of dried blood spots. Analysis of cord blood samples is undertaken as a clinician-led request rather than for general screening. If umbilical cord blood samples are used, they can be examined by haemoglobin electrophoresis using cellulose acetate at alkaline pH or citrate agar at acid pH,²³ or by HPLC or IEF. If any abnormality is detected, a confirmatory technique should also be undertaken.

Babies provisionally diagnosed as having Hbs SS, SC, SD^Punjab, SO^Arab or Sβ thalassaemia should be retested within 6–8 weeks of birth. After confirmation of the diagnosis, they should be followed in a paediatric clinic, immediately started on prophylactic penicillin to prevent pneumococcal infections, and appropriately managed in the long term.³ β thalassaemia major is suspected when Hb A is either absent or greatly reduced at birth. Such babies are retested for confirmation. The diagnosis of β thalassaemia trait cannot be reliably made until 12 months of age unless DNA techniques are used (see p. 146).

Detection of an unstable haemoglobin

Haemoglobin variants exhibit a wide range of instability but the clinically unstable haemoglobins can be detected by both the heat stability test and the isopropanol test.³⁹ However, minor degrees of instability that have little or no clinical significance may need other techniques. The unstable haemoglobins are frequently silent using electrophoretic or chromatographic techniques and tests for haemoglobin instability are essential in the detection or exclusion of an unstable haemoglobin.

Several methods are available for the demonstration of haemoglobin instability. Samples analysed should be as fresh as possible and certainly less than 1 week old. Controls should be of the same age as the test sample; a normal cord blood sample can be used as a positive control. The isopropanol test uses chemically prepared controls.

Heat Stability Test

Principle

When haemoglobin in solution is heated, the hydrophobic van der Waals bonds are weakened and the stability of the molecule is decreased.^40,⁴¹ Under controlled conditions, unstable haemoglobins precipitate, whereas stable haemoglobins remain in solution.

Reagent

Tris-HCl buffer, pH 7.4, 0.05 M. Tris, 6.05 g; water to 1 litre. Adjust the pH to 7.4 with concentrated HCl. (It is essential that Tris-sensitive electrodes are used.)

Method

1. Add 0.2 ml of lysate, freshly prepared by the purified haemolysate method (given on p. 309), to a tube containing 1.8 ml of buffer. The negative control is obtained from a fresh normal sample.

2. Place the tubes in a waterbath at 50°C. Examine the tubes at 60, 90 and 120 min for precipitation.

Interpretation and Comments

A major unstable haemoglobin will have undergone marked precipitation at 60 min and profuse flocculation at 120 min. The normal control may show some (fine) precipitation at 60 min, but this should be minimal.

Isopropanol Stability Test

Principle

When haemoglobin is dissolved in a solvent such as isopropanol, which is less polar than water, the hydrophobic van der Waals bonds are weakened and the stability of the molecule is decreased. Under controlled conditions, unstable haemoglobins precipitate, whereas stable haemoglobins remain in solution. This method has the advantage that it does not require a 37°C waterbath and positive controls can be made by modification of the reagent buffer.⁴²

Reagents

Tris-HCl buffer, pH 7.4, 0.1 mol/l. Tris, 12.11 g; water to 1 litre. Adjust the pH to 7.4 with concentrated HCl. It is essential that (Tris-sensitive electrodes are used.)

Isopropanol buffer, 17%. Make 17 volumes of isopropanol up to 100 volumes with tris-HCl buffer. The 17% isopropanol buffer solution may be stored in a tightly stoppered glass bottle for 3 months at 4°C.

Positive controls. These are buffers produced by adding small amounts of zinc to the standard 17% isopropanol buffer. For the strongest positive control (5+), add 0.6 mmol/l zinc acetate and for the weaker positive control (1+), add 0.1 mmol/l zinc acetate to the buffer. Samples containing haemoglobin E or haemoglobin F can also be used as weak positive controls.

Method

1. Prepare oxyhaemoglobin haemolysates from test and normal control samples as given on p. 309.

2. Pipette 2.0 ml of the standard isopropanol buffer into two tubes, followed by 2.0 ml of the 1+ and 5+ control solutions, respectively, into two further tubes.

3. Add 0.2 ml of test sample to the first tube. Add 0.2 ml normal control sample into the three remaining tubes.

4. Place the tubes in a waterbath at 37°C for 30 min. Examine the tubes at 5, 20 and 30 min for turbidity and fine flocculation.

Interpretation and Comments

A normal sample will remain clear until 30 min, when a slight cloudiness may appear. Some unstable haemoglobins will show clearly observable precipitation even after 5 min incubation, whereas milder variants will not show precipitation until 20 min.

Positive results may be given by samples containing as little as 10% Hb F or by samples containing increased methaemoglobin as a result of prolonged storage. If the normal sample undergoes premature precipitation, check the temperature of the waterbath because it is likely to be higher than 37°C.

False-negative results should be avoided by continuing the incubation until the normal control undergoes precipitation.

Detection of Hb Ms

Methaemoglobin (Hi) has iron present in the ferric form. Inherited variants of haemoglobin that undergo oxidation to methaemoglobin more readily than Hb A are referred to as Hb Ms. This is one of the causes of a very rare condition, congenital methaemoglobinaemia. The other cause of inherited methaemoglobinaemia is methaemoglobin reductase deficiency (see p. 240). Methaemoglobin levels vary, but may be as high as 40% of the total haemoglobin. Methaemoglobinaemia per se may also be caused by oxidant chemicals.

Methaemoglobin variants may be detected by haemoglobin electrophoresis at pH 7, but almost all can be distinguished from methaemoglobin A (Hi A) by their absorption spectra. Each methaemoglobin has its own distinct absorption spectrum. Hi A has two absorption peaks at 502 nm and 632 nm, whereas the peak absorbances for the variant Hb Ms are at different wavelengths (Fig. 14.11).

Figure 14.11 Absorption maxima of methaemoglobins in the range of 450–650 nm. Normal methaemoglobin is shown by a solid line; Hb M Saskatoon is shown by a dotted line.

(Reproduced with permission from Lehmann H, Huntsman KG 1974 Man’s Haemoglobins, 2nd edition, p. 214. North-Holland, Amsterdam.)

Detection of altered affinity haemoglobins

Electrophoretic and chromatographic techniques are frequently unsuccessful in separating these abnormal haemoglobins and cannot be relied on for detection because the amino acid substitution often does not involve a change in charge.

The most informative investigation is the measurement of the oxygen dissociation curve (see p. 268). The most significant finding is a decreased Hill’s constant (‘n’ value) because this can only come about by a change in the structure of the haemoglobin. The p50 may be either increased (low-affinity haemoglobin) or decreased (high-affinity haemoglobin). High-affinity haemoglobins result in an increase in Hb level, whereas low-affinity haemoglobins result in a decrease in Hb level. The p50 alone may be affected by other factors such as the high concentration of 2,3-DPG in pyruvate kinase deficiency. (Aspects of this are discussed in Chapter 11.)

Differential diagnosis of common haemoglobin variants

Suggested methods for differential diagnosis are given in Table 14.6 and Figure 14.12 gives a comparison of some common variants using different techniques.

Table 14.6 Methods helpful in the differential diagnosis of common structural variants

Initial Finding on Cellulose Acetate Electrophoresis	Most Likely Variant	Differentiation
Band in position of Hb S	Hb S, D, G-Philadelphia, Lepore	Blood count, quantitation, solubility test, citrate agar/acid gel electrophoresis, IEF, HPLC
Band in position of Hb C	Hb C, E, O-Arab	Quantitation, citrate agar/acid gel electrophoresis, IEF, HPLC
Very fast band	Hb I, H	H bodies

IEF, isoelectric focusing; HPLC, high-performance liquid chromatography.

Figure 14.12 Comparison of the relative mobilities of some abnormal haemoglobins by different methods. The position of Hbs A, S and C and their corresponding chains are indicated by the vertical lines.

(Adapted from ICSH.²³)

Investigation of suspected thalassaemia

A suggested scheme of investigations is shown in Figure 14.13; the methods used are listed in the following.

1. Estimates of Hb A₂ between 3.3% and 3.8% need careful assessment and should be repeated.

2. Hb A₂ values in α thalassaemia trait are usually below 2.5%. Some types of β thalassaemia trait have normal Hb A₂ values.

Figure 14.13 Suggested scheme of investigation for thalassaemia.

Methods for Investigation of Thalassaemia

1. Full blood count with red cell indices and blood film and, in selected cases, reticulocyte count

2. Hb A₂ measurement by cellulose acetate electrophoresis with elution (see p. 322)

3. Hb A₂ measurement of microcolumn chromatography (see p. 323)

4. Automated HPLC (see p. 325)

5. Quantitation of Hb F (see p. 325 or HPLC)

6. Assessment of the distribution of Hb F (see p. 327)

7. Assessment of iron status (see p. 179)

8. Demonstration of red cell inclusion bodies (see p. 329)

9. DNA analysis (see p. 146).

Blood Count and Film

The blood count, including haemoglobin and red cell indices, provides valuable information useful in the diagnosis of both α and β thalassaemia. In classical cases, there will be an elevation in the red cell count, accompanied by a decrease in MCV and MCH. The MCHC and red cell distribution width (RDW) are often normal in thalassaemia trait, whereas in iron deficiency anaemia they are more likely to be abnormal. The blood film may show features such as target cells, basophilic stippling and microcytosis in the absence of hypochromia, which point to a diagnosis of thalassaemia trait. Anisochromasia, which is a feature of iron deficiency, is not usual in α and β thalassaemia trait, although it may be seen in haemoglobin H disease. Haemoglobin H disease is also characterized by marked poikilocytosis. The reticulocyte count is increased in haemoglobin H disease.

Quantitation of Hb A₂

An increased Hb A₂ level is characteristic of heterozygous β thalassaemia and its accurate measurement is required for the diagnosis or exclusion of β thalassaemia trait. Estimations may be made by elution after cellulose acetate electrophoresis or by chromatography, either microcolumn or HPLC.

Measurement of Hb A₂ by Elution from Cellulose Acetate

Principle

Haemolysate is separated into its component fractions by alkaline electrophoresis on cellulose acetate membrane. The relative proportions of the separated fractions are quantitated by spectrometry of the eluates of the separated fractions.^43,⁴⁴

Equipment

Electrophoresis tank and power pack. see p. 310

Wicks of double filter paper or chromatography paper

Cellulose acetate membranes (78 × 150 mm).

Reagent

TEB buffer, pH 8.5. Tris(hydroxymethyl) methylamine, 40.8 g; disodium EDTA, 2.4 g; orthoboric acid, 12.8 g; water to 4 litres.

Method

1. Prepare a purified haemolysate from washed red cells as described on p. 309. The haemolysate may be kept at 4°C for up to 1 week before analysis.

2. With the power supply disconnected, pour equal amounts of TEB buffer into both the anode and the cathode chambers. Cut lengths of filter or chromatography paper, soak them in the buffer chamber and place them along the bridge supports as wicks. Set the bridge gap to 7 cm.

3. Soak the cellulose acetate by carefully floating the cellulose acetate sheet onto the surface of the buffer, making sure that no air bubbles are trapped underneath it. When the sheet has absorbed the buffer, submerge the sheet. Leave for at least 5 min, remove and blot carefully between two sheets of blotting paper.

4. Position the cellulose acetate across the bridge supports so that the long end of the sheet is on the anodal side. Using a ruler as a guideline, apply 30 μl of lysate to each sheet 1 cm from the cathode in a single line. Leave 1 cm margin at each end of the application line.

5. Run at a constant voltage of 250 V until separation is complete. This will take approximately 60 min and there should be at least a 1 cm gap between the A and A₂ bands at the end of the run. Check the separation at 40 min, ensuring that the Hb A (or variant such as Hb H, Hb J or Hb N) does not travel onto the wicks. Check again at 5–10 min thereafter. Samples with a variant band between the A and A₂ bands will take up to 30 min longer to obtain satisfactory separation.

6. Remove the sheet, holding it carefully at the anodal wick contact; do not place on a work surface. Cut off and discard the cellulose acetate that has been in contact with the cathodal wick. Cut a ‘blank’ strip of approximately 2 cm wide from the cathodal end of one sheet (do not use cellulose acetate that includes the application line). Cut out the Hb A₂ section, cutting the band into pieces approximately 1 cm square directly into a clean universal container. Cut out any variant band in the same way and finally cut out the Hb A band.

7. Add 4 ml of distilled water to both the blank and Hb A₂ containers and add 16 ml water to the Hb A container. Variant bands are usually eluted in 8 ml of water, although this may vary.

8. Mix the eluates for 20 min and mix again by inversion just before measuring the absorbance.

9. Read the absorbance of the blank against water at 415 nm. This reading should be <0.005. Read the absorbances of the haemoglobin solutions at 415 nm against the cellulose acetate blank.

Calculation

Interpretation and Comments

For interpretation of results and normal ranges, see p. 325. Duplicate values obtained should be within 0.2%. This method is inaccurate in the presence of Hb C, Hb E and Hb O^Arab because they do not separate from Hb A₂.

The procedure is useful for the measurement of haemoglobin variants: in these cases, the volume of water used for elution should be adjusted to the apparent quantity of the variant as judged on electrophoresis. Particular care must be taken when cutting strips on which a variant of haemoglobin (e.g. Hb S) is present because the separation between Hb S and Hb A₂ is less certain.

To obtain accurate and precise results, use the same cuvette when reading the blank, Hb A₂ and Hb A absorbance of each sample. Read the blank, Hb A₂ and Hb A in that order to minimize the effects of carryover. Some types of cellulose acetate are unsuitable for elution; this can be detected by a very high blank reading. The haemoglobin concentration of the haemolysate is important: the absorbance reading of the haemoglobin A₂ must be at least 0.1 absorbance unit because low values will give inaccurately low Hb A₂ results.

Measurement of Hb A₂ by Microcolumn Chromatography

Principle

Microcolumn chromatography depends on the interchange of charged groups on the ion exchange cellulose with charged groups on the haemoglobin molecule. When a mixture of haemoglobins is adsorbed onto the cellulose, a particular haemoglobin component may be eluted from the column using a buffer (developer) with a specific pH and/or ionic strength, whereas other components (either a single haemoglobin or a mixture of haemoglobins) may be eluted by changing the pH or ionic strength of the developer. The separation of haemoglobin components depends on the pH and/or ionic strength of the developers used for the equilibration of the column and for the elution, the type of cellulose, the volume of the sample added, the size of the column, the gradient, flow rates and temperature. The following methods use the anion exchanger diethylaminoethyl (DEAE) cellulose (Whatman DE-52 microgranular pre-swollen), with Tris-HCl developers ⁴⁵ or glycine-KCN developers.⁴⁶

Measurement of Hb A₂ by Microcolumn Chromatography with Tris-HCl Buffers

Reagents

DE-52 ion exchange cellulose (Whatman)

Stock buffer 1.0 mol/l Tris. Tris, 121.1 g; water to 1 litre. It is essential that Tris-sensitive electrodes are used

Working buffer 1. KCN, 200 mg; stock buffer, 100 ml; water to 2 litres; adjust to pH 8.5 with concentrated HCl

Working buffer 2. KCN, 200 mg; stock buffer, 100 ml; water to 2 litres; adjust to pH 8.3 with concentrated HCl

Working buffer 3. KCN, 200 mg; stock buffer, 100 ml; water to 2 litres; adjust to pH 7.0 with concentrated HCl.

Important: If the buffers are stored at 4°C, they must be allowed to come to room temperature before use.⁴⁵

Method

1. Prepare the slurry by adding 10 g of DE-52 to 200 ml of buffer 1. Mix gently and allow the cellulose to settle. Decant the supernatant and add a further 200 ml of buffer 1, mix gently for 10 min, then adjust the pH of the thoroughly suspended cellulose to 8.5 with concentrated HCl. Allow the cellulose to settle, remove the supernatant and resuspend in a further 200 ml of buffer 1. Mix gently for 10 min and ensure the pH is 8.5. Allow to settle and remove enough buffer so that the settled cellulose constitutes about half the total volume.

2. Secure short-form pipettes vertically in a support rack. Place either a 3 mm glass bead or a small piece of cotton wool in the tapered part of the pipette to act as a support for the slurry.

3. Fill the pipettes with thoroughly suspended cellulose slurry and allow the column to pack to a height of 5–6 cm.

4. Dilute 1 drop of haemolysate (100 g/l) with 5 drops of buffer 1.

5. When the excess buffer has drained from the column, gently apply the diluted lysate to the top of the column and allow it to be adsorbed onto the resin. Do not allow the surface of the column to dry out.

6. Apply 8 ml buffer 2 gently to the column with a 10–15 cm length of polythene tubing attached to the top of the pipette acting as a reservoir. Collect the eluate in a 10 ml flask and make the volume up to 10 ml with buffer 2.

7. Elute the remaining Hb A, using 10 ml of buffer 3; collect the eluate and make the volume up to 25 ml with the remaining buffer 3.

8. Read the absorbance of the eluted haemoglobins at 415 nm in a spectrometer, using water as a blank.

Calculate the Hb A₂ as follows:

Interpretation and Comments

For interpretation and normal ranges, see p. 325. The technique is inappropriate in the presence of haemoglobin variants (see below). Factors affecting quality assurance include the concentration of haemoglobin applied to the column – excess haemoglobin will cause contamination of the Hb A₂ fraction with Hb A. An inadequate amount of haemoglobin will result in an eluate with an absorbance too low for accurate measurement.

The flow rate of the column may be adjusted by altering the height of the reservoir above the column. A flow rate of 10–20 ml/h is satisfactory. Raising the reservoir increases the flow rate but broadens the Hb A₂ band on the column, which will not affect quantitation providing there is adequate separation. To elute the Hb A₂ band, 8 ml of buffer 2 should be used; the greater part of that should elute between 4 and 6 ml.

Measurement of Hb A₂ by Microcolumn Chromatography with Glycine-Potassium Cyanide Developers

The method described as follows is suitable for samples containing variants such as Hb S. The elution of Hb A₂ is dependent on the pH of the ion exchanger and on the molarity of the developer.⁴⁶

Reagents

Developer A. Glycine, 15.0 g; KCN, 0.1 g; water to 1 litre

Developer B. NaCl, 9.0 g; water to 1 litre

DE-52 ion exchange cellulose (Whatman).

Method

1. Prepare the slurry by adding 50 g of DE-52 to 250 ml of developer. Mix gently, then allow to settle and remove the supernatant. Repeat this process at least twice, then adjust the pH of the thoroughly suspended cellulose to 7.6 with 0.1 mol/l HCl. If the slurry is made too acidic, it should be discarded because any attempt to readjust it would increase the total ionic concentration and therefore alter the elution pattern. The slurry may be stored for up to 4 weeks, but the pH should be checked and, if necessary, readjusted before use.

2. Secure short-form pipettes vertically in a support rack. Place either a 3 mm glass bead or a small piece of cotton wool in the tapered part of the pipette to act as a support for the slurry.

3. Fill the pipette with thoroughly suspended DE-52 slurry and allow the column to pack under gravity to a height of about 6 cm.

4. Check each batch of columns with a Hb AS haemolysate. The Hb A₂ should elute in the first 3–4 ml and the Hb S should elute in the next 15–20 ml of the developer.

5. Dilute 1 drop of lysate (100 g/l) with 6 drops of water.

6. When all the excess buffer has drained from the column, gently apply the diluted lysate to the top of the column and allow it to be adsorbed onto the resin. Do not allow the surface of the column to dry out.

7. Apply developer A gently to the column with a piece of polythene tubing attached to the top of the pipette acting as a reservoir. About 3–4 ml of developer should be used to elute the Hb A₂ band. Collect the eluate in a 5 ml flask and make the volume up to 5 ml with developer A.

8. Elute the remaining Hb A or Hb S + Hb A, using 15–20 ml of developer B; collect the eluate and make the volume up to 25 ml with developer B. If, at any stage, the flow through the column stops, it should be discarded.

9. Read the absorbance of the eluted haemoglobins at 415 nm in a spectrometer, using water as a blank.

Calculate the Hb A₂ as follows:

Modification for the Measurement of Hb S

To estimate the percentage of Hb S and the remaining haemoglobin as well as that of Hb A₂, Hb A₂ is eluted in the first 3–4 ml with developer A, Hb S is eluted in the next 15–20 ml of the same developer A, and the remaining haemoglobin is eluted with developer B. The eluate containing Hb A₂ is diluted to 5 ml and the eluates containing Hb S and the remaining haemoglobin are diluted to 25 ml. To ensure elution of all the Hb A₂ in the first 3–4 ml and all the Hb S in the next 15–20 ml, the pH of the ion exchanger may need adjustment following a test chromatogram.⁴⁶

Interpretation and Comments

Hb A₂ percentages tend to be very slightly lower using the Tris buffer system, but with either procedure there should be a distinction between normal and classical β thalassaemia trait subjects.⁴⁵ An advantage of the glycine-KCN method is less sensitivity to minor changes in the pH of the developer; also it may be used for samples containing Hb S.

It should be noted that measurement of Hb A₂ in the presence of Hb S is not usually a very useful test. It is not necessary in order to distinguish sickle cell trait from sickle cell/β⁺ thalassaemia and is not always reliable in distinguishing sickle cell anaemia from sickle cell/β° thalassaemia because there is often interaction with α thalassaemia trait. In these circumstances, family studies can be extremely helpful.

Measurement of Hb A₂ by High-Performance Liquid Chromatography

The principle of HPLC has been explained on p. 314. When this technology is used as the primary method for detecting variant haemoglobins, simultaneous quantitation of Hb A₂ and Hb F means that it can replace three separate traditional methods: haemoglobin electrophoresis, quantitation of Hb A₂ and quantification of Hb F. Each laboratory should establish its own reference range for the quantitation of Hb A₂ by this method, which should be similar to published ranges. Because the quantitation of Hb A₂ may be inaccurate in the presence of certain variant haemoglobins, such as Hb E, Hb Lepore and Hb S,³² each chromatogram should always be inspected. Inspection will also permit identification of specimens with a split A₂ band as the result of heterozygosity for a δ chain variant. If the quantity of a haemoglobin with the retention time of Hb A₂ is higher than expected, an alternative technique should be applied to confirm its identity because a peak labelled as Hb A₂ can be Hb E or another haemoglobin that elutes with Hb A₂.

Interpretation of Hb A₂ values

Hb A₂ values should be interpreted in relation to a reference range established in each individual laboratory using blood samples from the local population with a normal Hb and red cell indices.^11,⁴⁷^–⁵⁰ The standard operating procedure for the relevant method should be strictly followed and 95% reference ranges should be determined. Ranges may differ slightly between methods and between laboratories. For example, in one of our laboratories the range determined for microcolumn chromatography was 2.2–3.3%, whereas in the other it was 2.3–3.5%. Technical variables affecting the range may include the use of packed cells rather than whole blood. Results obtained by HPLC analysis may be 0.1–0.2% higher than the results obtained by electrophoresis with elution. Once a reference range is determined, there is still a practical problem with borderline results, given that repeat estimates may vary by 0.1–0.2%. We recommend that Hb A₂ levels of 3.4–3.7% be regarded as borderline and that the assay should be repeated both on the same sample and on a fresh sample. There is also evidence that Hb A₂ is elevated in patients with HIV infection.^51,⁵²

When assays are being performed for genetic counselling, it can be useful to investigate the partner whenever borderline results are obtained.

The Hb A₂ percentage should be interpreted with knowledge of the Hb and red cell indices (Table 14.7).

Table 14.7 Interpretation of Hb A₂ values

Hb A₂ RANGE (%)	Interpretation
>7.0	Hb A₂ values of >7.0% are extremely rare
	Exclude a structural variant
	Repeat Hb A₂ estimation
	Rare β thalassaemia mutations
3.8–7.0	β thalassaemia trait, unstable haemoglobin
3.4–3.7	Severe iron deficiency in β thalassaemia trait
	Additional δ chain variant with β thalassaemia trait
	(Total A₂ must be measured)
	Interaction of α and β thalassaemia
	Rare β thalassaemia mutations
	Presence of Hb S, making accurate measurement difficult
	Interaction of α thalassaemia and Hb S
	Analytical error, including carryover; repeat analysis
2.0–3.3	Normal
	δβ thalassaemia (if Hb F elevated)
	Rare cases of
	β thalassaemia trait, including coexisting
	β and δ thalassaemia and coexisting
	β and α thalassaemia
	α thalassaemia trait
	α or δ chain variant (total A₂ must be measured)
<2.0	δβ thalassaemia (if Hb F elevated)
	α thalassaemia trait
	Hb H disease
	Additional δ chain variant present (total Hb A₂ must be measured)
	δ thalassaemia
	Iron deficiency

Quantitation of Hb F

Hb F may be estimated by several methods based on its resistance to denaturation at alkaline pH, by HPLC or by an immunological method.⁵³ Of the alkaline denaturation methods, that of Betke et al.⁵⁴ is reliable for small amounts (<10–15%) of Hb F, whereas for levels of more than 50% and in cord blood, the method of Jonxis and Visser⁵⁵ is preferable; however, this method is not reliable at levels of less than 10%.

Immunological methods have been devised to measure Hb F by immunodiffusion,⁵⁶ for which commercial kits are available (Helena Laboratories, Beaumont, Texas, USA) and by enzyme-linked immunoassay (ELISA).⁵⁷

Modified Betke Method for the Estimation of Hb F

Principle

To measure the percentage of Hb F in a mixture of haemoglobins,⁵⁴ sodium hydroxide is added to a lysate and, after a set time, denaturation is stopped by adding saturated ammonium sulphate. The ammonium sulphate lowers the pH and precipitates the denatured haemoglobin. After filtration, the quantity of undenatured (unprecipitated) haemoglobin is measured. The proportion of alkali-resistant (fetal) haemoglobin is then calculated as a percentage of the total amount of haemoglobin present.

Equipment

Filter paper. Whatman No. 42

Vortex mixer

Glass tubes.

Reagents

Cyanide solution. Potassium cyanide, 25 mg; potassium ferricyanide, 100 mg. Dissolve in 500 ml distilled water. Store in a dark bottle

Saturated ammonium sulphate solution. Bring 1 litre of water to the boil and add ammonium sulphate until the solution is saturated. Cool and equilibrate at 20°C before use

Sodium hydroxide solution, 1.2 mol/l. Sodium hydroxide 4.8 g; distilled water to 100 ml. Prepare monthly. Equilibrate at 20°C before use.

Method

1. Prepare a lysate as described on p. 309. The lysate may be stored at 4°C for up to 1 week before use.

2. Add 0.25 ml lysate to 4.75 ml cyanide solution to make a solution of haemiglobincyanide (HiCN).

3. Transfer 2.8 ml of the haemiglobincyanide solution to a glass test tube and allow to equilibrate at 20°C.

4. Rapidly add 0.2 ml of 1.2 mol/l of NaOH and mix on a vortex mixer for 2–3 s.

5. After exactly 2 min, rapidly add 2 ml saturated ammonium sulphate solution and mix on a vortex mixer. Leave tubes to stand for 5–10 min at 20°C.

6. Filter twice through the same Whatman No. 42 filter paper, using a clean test tube to collect the filtrate each time. If the filtrate is not completely clear, filter again through the same paper. This filtrate contains the alkali-resistant haemoglobin.

7. To measure the total haemoglobin, transfer 0.4 ml of the haemiglobincyanide solution from step 2 into another tube and add 13.9 ml of water.

8. Read the absorbance of the alkali-resistant and total haemoglobin at 420 nm against a water blank.

9. Calculate the percentage alkali-resistant haemoglobin as follows:

Interpretation and Comments

Elevation of Hb F has a variety of causes (see p. 307). In very exceptional situations, other abnormal haemoglobins will also exhibit resistance to alkali, giving high results. It is imperative that haemoglobin electrophoresis or HPLC is done on these samples tested for Hb F to exclude the possibility of an unusual variant being present.

A normal and a raised Hb F control should be tested with every batch of samples. The raised Hb F control should ideally contain between 5% and 15% Hb F and this can be prepared from a mixture of cord and adult blood. Each laboratory must verify its own normal range, which should not differ significantly from published values; for adults the range is 0.2–1.0%.

Zago et al.⁵⁸ reported variability in the capacity of different batches of filter paper to absorb haemoglobin from the filtrate, which caused low results. It is necessary to equilibrate the temperature of the reagents to 20°C and to control the reaction temperature to 20°C to obtain accurate and reproducible results.

Method of Jonxis and Visser

Principle

The increased resistance of Hb F to denaturation by alkali is detected by recording the change in absorption at 576 nm in each minute, caused by the addition of ammonium hydroxide.⁵⁵ At this wavelength, the absorption of oxyhaemoglobin differs from that of the alkali haemochromogen that is formed on denaturation.

When the logarithm of the percentage of haemoglobin remaining undenatured is plotted against time, a straight line is obtained. By extrapolation to time zero, the percentage of Hb F in the original sample can be calculated.

Reagents

Ammonium hydroxide solution. NH₄OH, 100 g; water to 1 litre

Sodium hydroxide solution, 0.06 mol/l. Sodium hydroxide, 2.4 g; water to 1 litre.

Method

1. All reagents should be allowed to reach room temperature before use. Add 0.1 ml of blood or lysate (100 g/l) to 10 ml of water and mix.

2. Add 2 drops of ammonium hydroxide solution and mix.

3. Measure the absorbance in a spectrophotometer at 576 nm (A_B).

4. Add 0.1 ml of the same blood or lysate to 10 ml of sodium hydroxide solution; then add 2 drops of ammonium hydroxide solution and mix thoroughly.

5. Measure the absorbance in a spectrometer at 576 nm at every minute for 15 min (A_T); then incubate the solution at 37°C for 15 min, cool to room temperature and measure the absorbance (A_E). The ratio A_B:A_E should be constant.

6. Calculate the percentage of undenatured haemoglobin at each minute as follows:

Plot the percentage on the logarithmic scale of semilogarithmic paper against time. This should produce a straight line from which the original amount of Hb F at time zero can be found by extrapolation.

Interpretation and Comments

Comments regarding controls and normal ranges given for the Betke method are also applicable to this method. In addition, the Jonxis and Visser method requires an accurate spectrometer because the maximum absorption peak at 576 nm is very narrow and the difference in extinction between oxyhaemoglobin and alkali haemochromogen is relatively small. For interpretation of results, see p. 307.

Radial Immunodiffusion

The radial immunodiffusion procedure ⁵⁶ can be used for the quantitation of Hb F. The principle is based on an antibody–antigen reaction; the anti-Hb F is incorporated into the gel support medium, resulting in the formation of a visible opaque precipitin ring.

The square of the diameter of this ring is directly proportional to the concentration of Hb F. A standard curve must be prepared from samples containing known levels of Hb F plotted against their haemoglobin concentrations. Helena Laboratories market a kit containing prepared plates, a microdispenser and a measuring device.

The method is simple, but the formation of the precipitin rings requires at least 18 h of incubation at room temperature. For this reason, rapid diagnostic work is not possible. Care must be taken with sample application because damage to the plate wells results in asymmetric precipitin rings and erroneous measurements.

Assessment of the intracellular distribution of Hb F

Differences in the intracellular distribution of Hb F are used to differentiate between heterozygotes for δβ thalassaemia and the classical African type of HPFH. In the former, it can be shown that not all red cells contain Hb F (heterocellular distribution), whereas in the latter every cell contains Hb F (pancellular distribution), although there is some variability in content from cell to cell. It has been suggested that a heterocellular distribution may be more apparent than real and merely reflects that high levels of Hb F tend to give a more pancellular distribution than lower levels. For this reason, results should be treated with caution and not used to make a diagnosis in isolation.

Two techniques have been widely used for demonstrating intracellular Hb F distribution. The most frequently used is the acid elution test of Kleihauer ⁵⁹ that was originally developed for the detection of fetal red cells in the maternal circulation following transplacental haemorrhage. This method is described on p. 338. Less frequently used is the more sensitive immunofluorescent technique described in the following.

Immunofluorescent Method

Principle

Anti-Hb F antibody binds specifically and quantitatively to Hb F in fixed red cells. These cells can be identified after treatment with a second fluorescent-labelled antibody directed against the anti-Hb F.¹⁴

Equipment

Glass slides

Coplin jars

Microscope. Equipped with accessories for ultraviolet (UV) fluorescence

Moist chamber. Made from a Petri dish with moistened filter paper in the bottom.

Reagents

Phosphate buffered saline (PBS), pH 7.1. see p. 622

Rabbit antihuman Hb F serum. Dilute the antiserum 1 in 64 in PBS; store in small aliquots at −20°C. Stable for several months

Sheep (or goat) antirabbit immunoglobulin labelled with fluorescein isothiocyanate. Dilute 1 in 32 in PBS; store in small aliquots at −20°C. Stable for several months

Fixative. Acetone, 90 ml; methanol, 10 ml.

Method

1. Prepare thin blood films and allow to dry overnight.

2. Fix for 5 min at room temperature, shake off excess fixative and rinse immediately in PBS. If the films are too thick, they will peel off at this stage.

3. Rinse the slides in water and allow to dry.

4. Layer 5 μl of the anti-Hb F antisera onto the slide.

5. Incubate in the moist chamber at 37°C for 30 min or at room temperature for 60 min.

6. Rinse the slides thoroughly in PBS to remove any unbound antiserum.

7. Rinse the slides in water and allow to dry.

8. Layer 5 μl of the antirabbit antisera onto the slide.

9. Incubate in the moist chamber at 37°C for 30 min or at room temperature for 60 min.

10. Rinse the slides thoroughly in PBS to remove any unbound antiserum.

11. Rinse the slides in water and allow to dry.

12. Examine microscopically using a ×40 objective and filters suitable for use with fluorescein isothiocyanate. To quantitate the number of Hb F-containing cells, count the total number of cells in a field under white light using an eyepiece grid, then the number of stained cells under the UV light. If the level of Hb F is less than 10%, at least 2000 cells should be counted.

Comments

In normal adults, from 0.1 to 7.0% of cells show detectable fluorescence. The proportion of positive cells correlates well with the percentage of Hb F as measured by alkali denaturation at levels between 0.5% and 5.0%. As little as 1 pg of Hb F per cell can be detected, giving much greater sensitivity than the acid elution method. This increased sensitivity, however, may make a heterocellular distribution appear pancellular if the proportion of Hb F is greater than 10%.⁵³

Interpretation of Hb F values

See Table 14.8.^11,¹⁴

Table 14.8 Interpretation of Hb F values

Hb F range (%)	Interpretation
0.2–1.0	Normal results
1.0–5.0	In approximately 30% of β thalassaemia traits
	Some heterozygotes for a variant haemoglobin
	Some homozygotes for a variant haemoglobin
	Some compound heterozygotes for a variant haemoglobin and β thalassaemia
	Some individuals with haematological disorders (aplastic anaemia, myelodysplastic syndromes, juvenile myelomonocytic leukaemia)
	Some pregnant women (2nd trimester)
	Sporadically in the general population, particularly in Afro-Caribbeans (representing heterozygosity for non-deletional HPFH)
5.0–20.0	Occasional cases of β thalassaemia trait
	Some homozygotes for a variant haemoglobin including some sickle cell anaemia^a
	Some compound heterozygotes for a variant haemoglobin and β thalassaemia
	Some types of heterozygous HPFH δβ thalassaemia
15.0–45.0	Heterozygous HPFH African type (usually >20%)
	Some cases of β thalassaemia intermedia
>45.0	β thalassaemia major
	Some cases of β thalassaemia intermedia
	Neonates
>95.0	Homozygous African-type (deletional) HPFH
	Some neonates (particularly if premature)

a Highest levels seen with Arab-Indian haplotype and in patients treated with hydroxycarbamide.

Assessment of iron status in thalassaemia

Concurrent iron deficiency makes the diagnosis of thalassaemia trait more difficult because it masks the typical blood picture and can reduce Hb A₂ synthesis.^47,^48,⁵⁰ In β thalassaemia trait, dependent on the severity of the anaemia, the Hb A₂ value may be reduced to borderline or even to normal levels (3.0–3.5%). However, in many patients with β thalassaemia trait and iron deficiency, the Hb A₂ will still be raised.

Whenever possible, individuals should not be investigated for the presence of thalassaemia trait if they are iron deficient. Iron stores are usually replete after 3–4 months of treatment with iron. However, if a pregnant woman is suspected of having a thalassaemia trait, it is not possible to wait for the correction of iron deficiency to establish the diagnosis. The woman and her partner should be tested without delay, with DNA analysis of globin genes being carried out if both are suspected of having thalassaemia trait (see Chapter 8, p. 146).

In addition to traditional methods for iron assessment, such as measurement of serum ferritin or serum iron plus total iron-binding capacity, estimation of zinc protoporphyrin (see pp. 192) is of potential value. This test can be carried out on an EDTA sample within a haematology laboratory and is a measure of iron incorporation at the cellular level.

Red cell inclusions

The most important red cell inclusions found in the haemoglobinopathies are Hb H inclusion bodies (precipitated β chain tetramers) found in α thalassaemia,⁶⁰ α chain inclusions found in β thalassaemia major^7,⁶¹ and Heinz bodies found in unstable haemoglobin diseases.^41,⁶²

Precipitated α chains are found in the cytoplasm of nucleated red cell precursors of patients with β thalassaemia major; they can be demonstrated by supravital staining of the bone marrow with methyl violet (as can Heinz bodies) and appear as irregularly shaped bodies close to the nucleus of normoblasts. After splenectomy they may also be found in the peripheral blood normoblasts and reticulocytes. Heinz bodies (insoluble denatured globin chains) form as a result of exposure to oxidant drugs or chemicals and develop spontaneously in glucose-6-phosphate dehydrogenase (G6PD) deficiency and in the unstable haemoglobin diseases. In unstable haemoglobin diseases, they are usually only seen in the peripheral blood after splenectomy but may be demonstrated in patients with an intact spleen if their blood is kept at 37°C for 24–48 h. The use of methyl violet and of brilliant cresyl blue in the demonstration of precipitated α chain and Heinz bodies is described on p. 336.

Demonstration of Hb H Inclusion Bodies

Reagent

Staining solution. 1.0% brilliant cresyl blue or New methylene blue. New batches of stain must be tested with a known positive control because the redox action of the dyes may vary from batch to batch.

Method

1. Mix 2 volumes of fresh blood (within 24 h of collection) with 1 volume of staining solution.

2. Incubate at 37°C for 2 h or at room temperature for 4 h.

3. Resuspend the cells and spread a thin blood film.

4. Examine the film as for a reticulocyte count. The inclusion bodies appear as multiple greenish-blue dots, like the pitted pattern on a golf ball (see p. 337). They can be readily distinguished from reticulocytes, which exhibit uneven reticular material or infrequent fine dots.

Interpretation and Comments

In α⁺ thalassaemia trait, only a very occasional H body (1:1000 to 1:10 000) is usually seen; they are more numerous in α° thalassaemia, but the number of cells developing inclusions is not reliable in differentiating the various gene deletion patterns seen in α thalassaemia and the absence of demonstrable inclusions does not preclude a diagnosis of α thalassaemia trait. This test is most useful in Hb H disease, where inclusions are usually found in more than 30% of red cells.

Fetal diagnosis of globin gene disorders

Prenatal diagnosis of globin gene disorders ²¹ is carried out if the fetus is at risk of thalassaemia major or a severe form of sickle cell disease such as sickle cell anaemia. Two approaches to fetal diagnosis are available: globin chain synthesis (used if the putative father is not available) and DNA analysis. DNA can be obtained from a chorionic villus sample or from amniotic fluid. Methods used for DNA analysis are described in Chapter 8.

When a potentially at-risk couple is detected, they will require counselling, and if a fetal diagnosis is requested, it is necessary to confirm the parental haemoglobin phenotype. The family or parental blood samples are sent to the diagnostic centre and the timing of fetal sampling is arranged.