If haemolytic anaemia is suspected, a full blood count, reticulocyte count and blood film should always be performed. The blood count shows a reduced haemoglobin concentration (Hb) and, usually, an increased mean cell volume (MCV). The increased MCV is attributable to the fact that reticulocytes, which may constitute a significant proportion of total red cells, are larger than mature red cells. The abnormalities that may be detected in the blood film and their possible significance in acquired haemolytic anaemia are shown in Table 13.1. Abnormalities detected in the blood film will direct further investigations. For example, a Heinz body preparation would be relevant if irregularly contracted cells were present, particularly if there appeared to be red cell inclusions. Similarly, a direct antiglobulin test (DAT) would be indicated if the blood film showed spherocytes. Various inherited forms of haemolytic anaemia enter into the differential diagnosis of suspected acquired haemolytic anaemia. Thus, spherocytes could be attributable to hereditary spherocytosis as well as to autoimmune or alloimmune haemolytic anaemia. Haemolysis with irregularly contracted cells could be attributable not only to oxidant exposure but also to an unstable haemoglobin, homozygosity for haemoglobin C or glucose-6-phosphate dehydrogenase (G6PD) deficiency.

Table 13.1 Abnormalities that may be detected on blood film examination and their possible significance

Morphological abnormality observed on blood film examination	Type of acquired haemolytic anaemia suggested
Schistocytes	Fragmentation syndromes including microangiopathic haemolytic anaemia and mechanical haemolytic anaemia
Spherocytes	Autoimmune, alloimmune or drug-induced immune haemolytic anaemia, paroxysmal cold haemoglobinuria, burns, Clostridium perfringens sepsis
Microspherocytes	Burns, fragmentation syndromes
Irregularly contracted cells	Oxidant damage, Zieve’s syndrome
Ghost cells, hemi-ghosts and suspicion of Heinz bodies	Acute oxidant damage
Marked red cell agglutination	Cold-antibody-induced haemolytic anaemia
Minor red cell agglutination	Warm autoimmune haemolytic anaemia, paroxysmal cold haemoglobinuria
Red cell agglutination plus erythrophagocytosis	Particularly characteristic of paroxysmal cold haemoglobinuria
Hypochromia, microcytosis and basophilic stippling	Lead poisoning
Erythrophagocytosis	Paroxysmal cold haemoglobinuria
Atypical lymphocytes	Cold-antibody-induced haemolytic anaemia associated with infectious mononucleosis or, less often, other infections
Lymphocytosis with mature small lymphocytes and smear cells	Autoimmune haemolytic anaemia associated with chronic lymphocytic leukaemia
Thrombocytopenia	Autoimmune haemolytic anaemia (Evans’ syndrome), thrombotic thrombocytopenic purpura, microangiopathic haemolytic anaemia associated with disseminated intravascular coagulation, paroxysmal nocturnal haemoglobinuria
Neutropenia	Paroxysmal nocturnal haemoglobinuria
No specific red cell features	Paroxysmal nocturnal haemoglobinuria

Immune haemolytic anaemias

Acquired immune-mediated haemolytic anaemias are the result of autoantibodies to a patient’s own red cell antigens or alloantibodies in a patient’s circulation, either present in the plasma or completely bound to red cells (e.g. transfused or neonatal red cells). Alloantibodies may be present in a patient’s plasma and react with antigens on transfused donor red cells to cause haemolysis. Alloantibodies may also occur in maternal plasma and cause haemolytic disease of the newborn. Autoimmune haemolytic anaemia (AIHA) may be ‘idiopathic’ or secondary, associated mainly with lymphoproliferative disorders and autoimmune diseases, particularly systemic lupus erythematosus. AIHA may also follow atypical (Mycoplasma pneumoniae) pneumonia or infectious mononucleosis and other viral infections. AIHA has also been reported following allogeneic bone marrow transplantation ¹ and other hematopoietic stem cell transplantation in both adult² and paediatric patients.³ Paroxysmal cold haemoglobinuria (PCH) also belongs to this group of disorders. Occasionally, drugs may give rise to a haemolytic anaemia of immunological origin that closely mimics idiopathic AIHA both clinically and serologically. This was a relatively common occurrence with α-methyldopa, a drug that is now used very infrequently, but it also occurs occasionally with other drugs. A larger range of drugs give rise to an antibody that is directed primarily against the drug and only secondarily involves the red cells. This is an uncommon occurrence. Such drugs include penicillin, phenacetin, quinidine, quinine, the sodium salt of p-aminosalicylic acid, salicylazosulphapyridine and cephalosporins.⁴

Types of Autoantibody

The diagnosis of an AIHA requires evidence of anaemia and haemolysis and demonstration of autoantibodies attached to the patient’s red cells (i.e. a positive DAT, see p. 279). A positive DAT may also be caused by the presence of alloantibodies (e.g. owing to a delayed haemolytic transfusion reaction), so details of any transfusion in the past months must be sought.

Autoantibodies can often be demonstrated free in the serum of a patient suffering from an AIHA. The ease with which the antibodies can be detected depends on how much antibody is being produced, its affinity for the corresponding antigen on the red cell surface and the effect that temperature has on the adsorption of the antibody, as well as on the technique used to detect it. The autoantibodies associated with AIHA can be separated into two broad categories depending on how their interaction with antigen is affected by temperature: warm antibodies, which are able to combine with their corresponding red cell antigen readily at 37°C and cold antibodies, which cannot combine with antigen at 37°C but form an increasingly stable combination with antigen as the temperature falls from 30–32°C to 2–4°C.

Cases of AIHA can similarly be separated into two broad categories according to the temperature characteristics of the associated autoantibodies: warm-type AIHA and the less frequent cold-type AIHA. The relative frequency of the two categories is illustrated in Table 13.2.⁵ In unusual instances, both warm autoantibody and cold autoantibody are detected in the patient’s serum and those cases are referred to as mixed-typed AIHA. This can be further classified into idiopathic or secondary, the latter often associated with systemic lupus erythematosus or lymphoma.^6,⁷

Table 13.2 Relative incidence of different types of autoimmune haemolytic anaemia ⁵

Warm Autoantibodies

The most common type of warm autoantibody is an immunoglobulin (Ig) G, which behaves in vitro very similarly to an Rh alloantibody; indeed, many IgG autoantibodies have a mimicking Rh specificity. IgA and IgM warm autoantibodies are much less common and when present they are usually formed in addition to an IgG autoantibody (Table 13.3).⁸

Table 13.3 Direct antiglobulin test in warm-antibody autoimmune haemolytic anaemia: incidence of different reactions to specific antiglobulin sera ⁸

Frequently, patients with warm-type AIHA have complement adsorbed onto their red cells and the red cells are therefore agglutinated by antisera specific for complement or a complement component such as C3d (Table 13.3). In these cases, the complement is probably not being bound by an IgG antibody but is on the cell surface as the result of the action of small and otherwise undetected amounts of IgM autoantibody.

IgG can fix complement and sometimes patients with warm-type AIHA appear to have a positive DAT with complement components only on the red cell surface. Similar results (positive DAT with complement only) are seen in some patients with no evidence of increased red cell destruction, due to binding of circulating immune complexes to the red cells.

Warm autoantibodies free in the patient’s serum are best detected by means of the indirect antiglobulin test (IAT) or by the use of enzyme-treated (e.g. trypsinized or papainized) red cells. (Antibodies that agglutinate unmodified cells directly in vitro are seldom present.) Not infrequently, antibodies that agglutinate enzyme-treated cells, sometimes at high titres, are present in the sera of patients in whom the IAT using unmodified cells is negative (Table 13.4). Occasionally, too, they are present in the sera of patients in whom the DAT is negative.

Table 13.4 Results of testing for free autoantibodies in the sera of 210 patients with warm-antibody autoimmune haemolytic anaemia ⁵

Antibodies in serum that can be shown to lyse (rather than simply agglutinate) unmodified red cells at 37°C in the presence of complement (warm haemolysins) are rarely demonstrable. If they are present, the patient is likely to suffer from extremely severe haemolysis. Antibodies in serum that lyse as well as agglutinate enzyme-treated cells but do not affect unmodified cells are, however, quite common. Their specificity is uncertain – they are not anti-Rh – and their presence is not necessarily associated with increased haemolysis.

Cold Autoantibodies

Cold autoantibodies are nearly always IgM in type. In vivo the majority do not cause haemolysis, although a minority can cause chronic intravascular haemolysis, the intensity of which is characteristically influenced by the ambient temperature. The resultant clinical picture is generally referred to as the cold haemagglutinin syndrome or disease (CHAD). Haemolysis results from destruction of the red cells by complement that is bound to the red cell surface by the antigen–antibody reaction, which takes place in the blood vessels of the exposed skin where the temperature is 28–32°C or less. The cold autoantibody in CHAD is monoclonal because this syndrome is the result of a low-grade lymphoproliferative disorder.

The red cells of patients suffering from CHAD characteristically give positive antiglobulin reactions only with anticomplement (anti-C′) sera. (The C′ notation is used to distinguish anticomplement antibodies from anti-C antibodies of the Rh system.) This is because of the presence of red cells that have irreversibly adsorbed sublytic amounts of complement; it is an indication of an antigen–antibody reaction that has taken place at a temperature below 37°C. The complement component responsible for the reaction with anti-C′ sera is the C3dg derivative of C3 (see p. 494).

In vitro, a cold-type autoantibody will often lyse normal red cells at 20–30°C in the presence of fresh human complement, especially if the cell-serum mixture is acidified to pH 6.5–7.0; it will usually lyse enzyme-treated red cells readily in unacidified serum and agglutination and lysis of these cells may still occur at 37°C. Most of these cold-type autoantibodies have anti-I specificity (i.e. they react strongly with the vast majority of adult red cells and only weakly with cord-blood red cells). A minority are anti-i and react strongly with cord-blood cells and weakly with adult red cells. Rarely, the antibodies have anti-Pr or anti-M specificity and react with antigens on the red cell surface that are destroyed by enzyme treatment.

Combined Warm and Cold Autoantibodies

In approximately 7% of cases with AIHA, both warm IgG antibody and cold IgM autoantibody are simultaneously detected in the patient’s serum.^6,⁷ These cases are referred to as ‘combined warm and cold AIHA’ or mixed-type AIHA. The serological characteristics in these patients are the presence of IgM cold autoantibody with a high thermal amplitude (reacting at or above 30°C) in association with a warm IgG autoantibody. In some cases, high- titre cold agglutinins (>1024 at 4°C) were reported^9,¹⁰ and in others the cold agglutinin titre were reported as >64 at 4°C.^11,¹²

Another quite distinct, but rarely encountered, type of cold antibody is the Donath–Landsteiner (D–L) antibody. This is IgG and has anti-P specificity. The clinical syndrome the antibody produces is PCH.

PCH is caused by a biphasic IgG autoantibody, usually with anti-P specificity, and is commonly seen as an acute condition in children. This antibody binds to the red cells in the cold but activates complement and causes haemolysis on rewarming to 37°C. Cases may be idiopathic or can be secondary to acute viral infection in children. Other tests of value in the diagnosis of PCH are discussed on p. 287.

The DAT is positive for complement only. A negative antibody screen by the standard IAT at 37°C is a common finding in a suspected case of PCH because of the low thermal amplitude of the autoantibody. If the antibody investigation is carried out at a lower temperature in PCH cases, panreactive cold antibodies may be detected because the majority of autoantibodies show anti-P specificity with thermal amplitude range up to 15–24°C. Usually the antibody titre is low (<64), even when investigated at 4°C.

Some of the characteristics of IgG, IgM and IgA antibodies are listed in Table 13.5.

Table 13.5 Main characteristics of IgG, IgM and IgA autoantibodies

The clinical, haematological and serological aspects of the AIHAs have been summarized by Dacie ¹³ and others.¹⁴^–¹⁸

Methods of Investigation

Many of the methods used in the investigation of a patient suspected of suffering from AIHA are described in Chapter 21. Detailed description is given here of precautions to be taken when collecting blood samples from patients and of methods of particular value in the investigations.

Collection of Samples of Blood and Serum

To determine the true thermal amplitude or titre of cold agglutinins requires that the blood sample is collected and maintained strictly at 37°C until serum and cells are separated. This can be achieved by collecting venous blood (clotted and ethylenediaminetetra-acetic acid [EDTA]-anticoagulated samples) and keeping it warmed at 37°C – ideally in an insulated thermos, but usually, in practice, by placing the sample tube in a beaker containing water at 37°C.

The red cells are available for antibody elution and the serum can be examined for free antibody or other abnormalities. The clotted sample should then be centrifuged to separate the serum at 37°C (e.g. in an ordinary centrifuge into the buckets of which has been placed water warmed to 37–40°C). The EDTA sample is used for the DAT and other tests involving the patient’s red cells. If the autoantibody in a particular case is known to be warm in type, the blood may be separated at room temperature; otherwise, as already indicated, this should be carried out at 37°C. When samples are sent by post, it is best to send separately: (1) serum (separated at 37°C) and (2) whole blood added to acid–citrate–dextrose (ACD) or citrate–phosphate–dextrose (CPD) solution. Sterility must be maintained.

Storage of Samples

Samples of a patient’s blood, while keeping quite well in ACD or CPD at 4°C, are more difficult to preserve than normal red cells. In particular, if marked spherocytosis is present, considerable lysis develops on storage. However, satisfactory eluates can be made from washed red cells that are frozen at –20°C for weeks or months.

The patient’s serum should be stored at –20°C or below in small (1–2 ml) volumes. If complement is to be titrated and the titration is not performed immediately, the serum should be frozen as soon as practicable at −70°C or below.

Scheme for Serological Investigation of Haemolytic Anaemia Suspected to be of Immunological Origin

It is important to consider which are the most useful tests to carry out and the order in which they should be done. A suggested scheme has been set out in the form of answers to questions.¹⁹ Whereas some information may be helpful in classifying the type of AIHA, the single most important practical consideration is to determine whether, in addition to an autoantibody, there is any underlying alloantibody present. This should be identified before transfusion is undertaken to avoid a delayed haemolytic transfusion reaction that would compound existing haemolysis.

1. Are the patient’s red cells ‘coated’ by immunoglobulins or complement (indicating an antigen–antibody reaction)?

Perform a DAT using a polyspecific ‘broad-spectrum’ reagent, which contains both anti-IgG and anti-C′. (If the DAT is negative, it is unlikely, although not impossible, that the diagnosis is AIHA. See DAT-negative AIHA, p. 281)

2. If the DAT is positive, are immunoglobulins or complement adsorbed to the red cells?

Repeat the DAT using monospecific sera (see p. 500) (i.e. anti-IgG and anti-C3d).

3. If immunoglobulins are present on the red cells, is there antibody specificity?

Prepare eluates from the patient’s red cells. Test these later (see item 6).

4. What is the patient’s blood group?

Determine the patient’s ABO and RhD and Kell type. The Rh phenotype is particularly important in warm-type AIHA; other antigens must be determined if alloantibodies are to be differentiated from autoantibodies (see p. 507).

5. Is there free antibody in the serum? How does it react and at what temperatures and by what methods can it be demonstrated? Is there any underlying alloantibody present?

Screen the serum with two or three red cell suspensions suitable for routine pretransfusion antibody screening (see p. 528) looking for agglutination and lysis at 37°C by the IAT (see p. 500). If positive, identify the antibody using an antibody identification panel.

a. If an alloantibody is identified, blood lacking the corresponding antigen must be selected for transfusion.

b. If no alloantibody is identified in the serum or plasma, it is safe to assume there is no alloantibody present, unless the patient has been transfused in the last month; in the latter case, a red cell eluate is required because an alloantibody may be bound to the recently transfused cells and there may not be free antibody detectable in the serum/plasma.

c. If the autoantibody is pan-reacting (i.e. is reacting against all panel cells), antibody adsorption tests are needed to remove the autoantibody so as to identify any underlying alloantibody. If the patient has not had a transfusion within the last 3 months, a ZZAP autoadsorption test is appropriate (see p. 283). If the patient has had a transfusion within the last 3 months, differential alloadsorption tests are needed. However, if the patient has had a transfusion within the last month, an eluate is required, irrespective of results of adsorption tests.

6. If there is a warm autoantibody, what is the specificity of the autoantibody?

Test the serum also at 20°C against antibody-screening cells to show whether cold or warm antibodies or a mixture of the two, are present in the serum.

Test the eluate against the antibody identification panel of red cells by IAT and by using enzyme-treated red cells (see p. 498). Titration of autoantibody may be useful in the presence of a strong alloantibody.

Titrate the serum/plasma by the methods that have given positive results in the screening test using the same panel of red cells (see item 5a).

7. If there is a cold antibody:

a. Has the antibody any specificity? Is it an autoantibody or an alloantibody? What is its titre?

b. What is the thermal range of the antibody?

Test the serum/plasma against a panel of O cells, O cord cells and patient’s own cells at 20°C. If an autoantibody is found, titrate at 4°C with ABO-compatible adult (I) cells, cord blood (i) cells, the patient’s cells and adult (i) cells (if possible):

i. Determine the highest temperature at which autoagglutination of the patient’s whole blood takes place (see p. 287).

ii. Titrate the patient’s serum/plasma at 20°C, 30°C and 37°C with pooled O adult cells, O cord cells, patient’s own cells and the panel of cells described earlier. If there was any agglutination or lysis at 37°C in the screening test (item 5), titrate with the appropriate cells at this temperature.

iii. If PCH is suspected, carry out the direct and two-stage indirect Donath–Landsteiner tests (see p. 287).

8. Is there is a mixture of both warm autoantibody and cold autoantibody?

The diagnosis of mixed-type AIHA can only be made after appropriate characterization of the serum autoantibodies. The serological characteristic in these cases is the presence of a cold IgM antibody with a high thermal amplitude (reacting at or above 30°C) in association with a warm IgG autoantibody.^6,⁹

9. Is a drug suspected as the cause of the haemolytic anaemia?

a. If a penicillin-induced haemolytic anaemia is suspected, test for antibodies using cells pre-incubated with penicillin (see p. 290).

b. If haemolysis induced by other drugs is suspected, add the drug in solution to a mixture of the patient’s serum, normal cells and fresh normal serum (see p. 290). Look for agglutination of normal and enzyme-treated cells and use the IAT.

10. Are there any other serological abnormalities?

Consider carrying out the following tests: serum protein electrophoresis and quantitative estimation of immunoglobulins, estimation of complement, tests for antinuclear factor, a screening test for heterophile antibodies (infectious mononucleosis screening test) and a test for mycoplasma antibodies.

Detection of Incomplete Antibodies by Means of the Direct Antiglobulin (Coombs) Test

Szymanski et al.²⁵ used an AutoAnalyser and used Ficoll and polyvinylpyrrolidone (PVP) to enhance agglutination by an anti-IgG serum highly diluted (usually to 1 in 5000) in 0.5% bovine serum albumin. In this sensitive system, the strength of agglutination was positively correlated with the serum γ-globulin concentration, being subnormal in hypogammaglobulinaemia and supranormal in hypergammaglobulinaemia.²⁶

Usually, in patients with hypergammaglobulinaemia in whom the DAT is positive, attempts to demonstrate antibodies in eluates fail (i.e. eluates are non-reactive).^27,²⁸

9. Cross-reacting antiphospholipid antibodies adsorbing non-specifically on to red cell membrane and binding to phospholipid-dependent epitopes. Positive DAT as a result of antiphospholipid antibodies has been documented in patients with primary antiphospholipid syndrome and antiphospholipid syndrome with systemic lupus erythematosus.²⁹ It has also been described in healthy blood donors.³⁰

10. Incidental findings of positive DAT with no clear correlation between DAT and anaemia. A positive DAT is a common finding in patients with sickle cell disease because of abnormal amounts of IgG coating red cells. There was no correlation between the amount of IgG on the patient’s red cells and the severity of anaemia.³¹

There was an association demonstrated between an elevated blood urea nitrogen and a positive DAT.³² Urea may alter the red cell membrane and enhance non-specific IgG adsorption.³²

11. Sensitization in vitro if a sample other than EDTA is used. If, for instance, clotted or defibrinated normal blood is allowed to stand in a refrigerator at 4°C or even at room temperature, and the antiglobulin test is subsequently carried out, the reaction may be positive because of the adsorption of incomplete cold antibodies and complement from normal sera. Samples of blood taken into EDTA or ACD and subsequently chilled do not give this type of false-positive result because the anticoagulant inhibits the complement reaction.

12. False-positive agglutination may occur with a silica gel derived from glass.³³ Also, albeit rarely, the DAT has been positive with the blood of apparently perfectly healthy individuals (e.g. blood donors). Such occurrences have not been satisfactorily explained (see below).

Positive DATs in Normal Subjects

The occurrence of a clearly positive DAT in an apparently healthy subject is a rare but well-known phenomenon. Worlledge ²⁰ reported a prevalence in blood donors of approximately 1 in 9000. In a later report, Gorst et al.³⁴ estimated that the prevalence was approximately 1 in 14 000 with an increasing likelihood of a positive test with increasing age. Their report and subsequent reports,^25,³⁵ suggest that the finding of a positive DAT, using an anti-IgG serum, in an apparently healthy person is usually of little clinical significance and that, although overt AIHA may subsequently develop, this is infrequent. In some such individuals the DAT eventually becomes negative.

Positive DATs in Hospital Patients

In contrast to the rarity of positive DATs in healthy people, positive tests are much more frequent in hospital patients. Worlledge ²⁰ reported that the red cells of 40 out of 489 blood samples (8.9%) submitted for routine tests were agglutinated by anti-C′ sera. Only one sample was agglutinated by an anti-IgG serum and this had been obtained from a patient being treated with α-methyldopa. Freedman³⁶ reported a similar incidence – 7.8% positive tests with anti-C′ sera. Lau et al.³⁷ used anti-IgG sera only. The tests were seldom positive (0.9% positive out of 4664 tests). The probable explanation for the relatively high incidence of positive tests with anti-C′ sera is that the reaction is between anti-C′ antibodies and immune complexes adsorbed to the red cells.

False-Negative Antiglobulin Test Results

There are several causes of false-negative test results:

1. Failure to wash the red cells properly: the antisera may then be neutralized by immunoglobulins or complement in the surrounding serum or plasma (see p. 501).

2. Excessive agitation at the reading stage: this may break up agglutinates, leading to a false-negative result.

3. The use of impotent antisera so that weakly sensitized cells are not detected.

4. The use of antisera lacking the antibody corresponding to the subclass of immunoglobulin responsible for the red cell sensitization.

5. The presence of an antibody that is readily dissociable and is eluted in the washing process.

These phenomena are largely negated by the use of column agglutination technology.

DAT-Negative Autoimmune Haemolytic Anaemia

Most hospital blood banks use polyspecific ‘broad-spectrum’ AHG reagents for screening for diagnosis of AIHA. These reagents contain antibody to human IgG and the C3d component of human complement and have little activity against IgA and IgM proteins. The incidence of IgA-only warm AIHA has been reported as 0.2% to 2.7%,³⁸ and the diagnosis may be missed if such polyspecific AHG is used for the DAT screen. In approximately 2–6% of patients who present with the clinical and haematological features of AIHA, the DAT is negative on repeated testing.^20,^39,⁴⁰

Low-affinity IgG autoantibodies dissociate from the red cells during the washing phase if a tube technique is used, resulting in a negative DAT. Alternatively, there may be few IgG molecules coating the red cells and this number may fall below the threshold of detection, which is 300–4000 molecules per red blood cell if a tube technique is used. In such cases, a positive DAT may be demonstrated by a more sensitive technique, such as a column agglutination method, an enzyme-linked immunoabsorbant assay or flow cytometry.⁴¹^–⁴³

If polyspecific AHG is used and the DAT remains negative with clinical evidence of haemolysis, a more sensitive technique should be used for further investigation.⁴⁴

The DiaMed DAT gel card, which contains a set of monospecific AHG reagents (i.e. anti-IgG, -IgA, -IgM, -C3c, -C3d and an inert control) can be used. Because there is no washing phase, this permits the detection of low-affinity IgG, IgA and IgM antibodies. A gel card can also pick up the rare IgA-only autoimmune haemolytic anaemia. In warm-type AIHA the DAT may be positive with anti-IgG or anti-IgG plus anti-C3d. In cold–type AIHA the DAT may be positive with anti-IgM or anti-IgM plus anti-C3d and in mixed-type AIHA the DAT may be positive with anti-IgG, anti-IgM and anti-C3d.

Preparing and Testing a Concentrated Eluate

The eluate technique concentrates low levels of immunoglobulin present on the red cell surface so that antibody may then be detected by screening the eluate with group O red cells by the IAT. Elution techniques reverse or neutralize the binding forces that exist between the red cell antigens and the antibody coating the cells. This may be achieved by several techniques (e.g. heat or alterations to the pH).

Manual Direct Polybrene Test

The following method ⁴⁵ is modified from that of Lalezari and Jiang.⁴⁶ Polybrene is a polyvalent cationic molecule, hexadimethrine bromide, that can overcome the electrostatic repulsive forces between adjacent red cells, bringing the cells closer together. When low levels of IgG are present on the red cell surface, antibody linkage of adjacent red cells is enhanced. The Polybrene is then neutralized using a negatively charged molecule such as trisodium citrate. Sensitized red cells remain agglutinated after neutralization of the Polybrene. Unsensitized red cells will disaggregate after neutralization.

Reagents

Polybrene stock. 10% Polybrene in 9 g/l NaCl, pH 6.9 (saline).

Working Polybrene solution. Dilute the stock Polybrene solution 1 in 250 in saline.

Resuspending solution. 60 ml of 0.2 mol/l trisodium citrate added to 40 ml of 50 g/l dextrose.

Washing solution. 50 ml of 0.2 mol/l trisodium citrate in 950 ml of saline.

Low-ionic medium. 50 g/l dextrose containing 2 g/l disodium EDTA. Adjust the pH of half the batch to 6.4. Store the remainder at the original pH (approx. 4.9); use this to repeat tests that are negative using a low-ionic medium at pH 6.4.

Method

Ensure that all reagents are at room temperature.

Positive control

Dilute an IgG anti-D in normal group AB serum. Find a dilution that gives a positive result with papainized cells but is negative by the IAT on standard testing with group O, D-positive red cells (a dilution of 1 in 10 000 is often suitable).

Negative control

Normal group AB serum that fails to agglutinate papainized group O, D-positive red cells.

1. Wash the cells four times in saline and make 3–5% suspensions of test and normal group O RhD red cells in saline.

2. Set up three 75 × 10 mm tubes as shown in Table 13.6. Leave at room temperature for 1 min.

3. Add 1 drop of working Polybrene solution to each tube and mix gently. Leave for 15 s at room temperature.

4. Centrifuge for 10 s at 1000 g. Decant, taking care to remove all the supernatant.

5. Leave for 3–5 min at room temperature before adding 2 drops of resuspending solution and mixing gently. Within 10 s aggregates will dissociate, leaving true agglutination in the positive tubes.

6. Read macroscopically after 10–60 s. Check all negative results microscopically and compare with the negative control.

7. Repeat negative tests using low-ionic medium at the lower pH (about 4.9).

Table 13.6 Setting up a direct manual Polybrene test

If the direct Polybrene test is negative, a supplementary antiglobulin test may be performed by washing the cells twice in the washing solution and testing with an anti-IgG antiglobulin reagent.

Determination of the Blood Group of a Patient with AIHA

ABO Grouping

No difficulty should be encountered in ABO grouping patients with warm-type AIHA using monoclonal reagents, but the presence of cold agglutinins may cause difficulties. The cells should in all cases be washed in warm (37°C) saline. They should then be groupable without any problem; the reactions must, however, be controlled with normal AB serum. Reverse grouping should be performed strictly at 37°C. Warm the known A₁, B and O cells to 37°C before adding them to the patient’s serum at 37°C. Read the results macroscopically.

RhD Grouping

When the DAT is positive, monoclonal anti-D reagents should be used; if cold agglutinins are present, perform the test at 37°C. Appropriate controls should be included (see p. 524).

Demonstration of Free Antibodies in Serum

The sera of patients suffering from AIHA often contain free autoantibodies. However, free autoantibody is also found with no haemolysis. As a result of improved reagent sensitivity, any clinically significant IgG complement-binding antibodies will be detected by current antibody screening methods.

Identification by Adsorption Techniques of Coexisting Alloantibodies in the Presence of Warm Autoantibodies

Adsorption techniques for the detection of alloantibodies present in the sera or eluates of patients with suspected or proved AIHA can be helpful in the following situations:

1. In screening for coexisting alloantibodies in patients with AIHA who have been pregnant or previously transfused and are found to have a panreactive antibody in their serum

2. In differentiating between autoantibodies and alloantibodies in the eluate of recently transfused patients with AIHA

3. In investigating haemolytic transfusion reactions owing to red cell alloantibodies in patients with AIHA.

In some cases of AIHA, an underlying alloantibody may be detected by titrating the patient’s serum and eluate against a panel of phenotyped reagent red cells. However, a high-titre autoantibody may mask the alloantibody; hence the need for adsorption techniques, especially in the situations outlined earlier.

Use of ZZAP Reagent in Autoadsorption Techniques

‘ZZAP’ reagent is a mixture of dithiothreitol and papain.⁴⁷ It dissociates an autoantibody already coating the patient’s red cells and enzyme treats the cells, thus increasing the amount of autoantibody that can subsequently be adsorbed onto the patient’s cells in vitro.

Reagents

Dithiothreitol (DTT). 0.2 mol/l.

Papain. 1%.

Phosphate-buffered saline (PBS). pH 6.8–7.2.

Prepare a suitable volume of ZZAP by making up the reagents in the following ratio: 0.2 mol/l DTT 5 volumes; and 1% papain 1 volume.

Check the pH and adjust to pH 6.0–6.5 using one drop at a time of 0.2 mol/l HCl or 0.2 mol/l NaOH.

Method

1. Add 2 volumes of ZZAP to 1 volume of packed red cells that have been washed four times. Incubate at 37°C for 30 min, mixing occasionally.

2. After incubation, wash the cells four times in saline, packing hard after the last wash.

3. Divide the cells into two equal volumes. To one volume, add an equal volume of the serum to be adsorbed. Incubate at 37°C for 1 h.

4. Centrifuge at 1000 g. Remove the serum and add to the remaining volume of cells.

5. Repeat the adsorption procedure.

6. Remove the adsorbed serum and store at –20°C or below for alloantibody screening or cross-matching, which may be performed by standard techniques.

Notes

The autoadsorption techniques should only be used in the following circumstances:

1. When the patient has not had a transfusion in the previous 3 months because the presence of transfused red cells may allow the adsorption of alloantibody as well as autoantibody

2. When at least 2–3 ml of packed red cells are available from the patient

3. When the autoantibodies present react well with enzyme-treated red cells. If they do not, heat elution should be substituted for ZZAP treatment. Heat elution may be performed by shaking the washed cells for 5 min in a 56°C waterbath and then washing the cells.

Alloadsorption Using Papainized R₁R₁, R₂R₂ and rr Cells

The method of alloadsorption using papainized R₁R₁, R₂R₂ and rr cells may be used when autoadsorption is not appropriate – for instance, when the patient has had a transfusion in the previous 3 months or when less than 2–3 ml of the patient’s red cells are available.

1. Select three group O antibody screening cells, which individually lack some of the blood-group antigens that commonly stimulate the production of clinically significant antibodies (e.g. c, e, C, E, K, Fy^a, Fy^b, Jk^a, Jk^b, S, s) (Table 13.7).

2. Papainize 2 ml of packed cells from each sample after washing the cells in saline four times.

3. Add to 1 ml of each sample of washed, packed, papainized cells, 1 ml of the patient’s serum. Incubate for 20 min at 37°C.

4. Centrifuge to pack the cells. Remove the supernatant serum and add it to the second 1 ml volume of papainized cells. Incubate for 1 h at 37°C.

5. Centrifuge again to pack the cells. Remove the supernatant and store at –20°C or below for further testing (e.g. alloantibody screening and cross-matching).

Table 13.7 Testing an alloabsorbed serum against a phenotyped panel of red cells

Method for Testing Alloadsorbed Sera

For alloantibody screening, each adsorbed serum is tested against a panel of phenotyped red cells by the IAT. For cross-matching, each adsorbed serum must be tested separately against the donor red cells by the IAT, using undiluted serum.

Example of alloantibody detection using the alloadsorption technique in a recently transfused patient with AIHA

The patient’s serum when first tested against a panel of group O phenotyped red cells revealed only panreactive antibodies. In contrast, three absorbed sera, A, B and C, obtained by adsorbing the patient’s serum with three selected phenotyped samples of group O cells, were shown to contain anti-E and anti-Jk^a when tested against a panel of phenotyped group O cells using the IAT. The results of testing the adsorbed sera, A, B and C, are shown in Table 13.7. The patient’s red cell phenotype was R₁r Jk (a− b−).

Explanation of the results of testing alloadsorbed sera, A, B and C

1. Because the R₁R₁-adsorbing cells were negative for the E and Jk^a antigens, adsorbed serum A could contain anti-E and anti-Jk^a. Testing the adsorbed serum A against the panel of cells suggested that this was the case.

2. Because the R₂R₂-adsorbing cells were positive for the E antigen but negative for the Jk^a antigen, adsorbed serum B could contain anti-Jk^a but not anti-E. Testing adsorbed serum B against the panel of cells confirmed the presence of anti-Jk^a.

3. Because the rr-adsorbing cells were negative for the E antigen but positive for the Jk^a antigen, adsorbed serum C could contain anti-E but not anti-Jk^a. Testing adsorbed serum C against the panel of cells confirmed the presence of anti-E.

4. Because the phenotype of the patient’s own red cells was R₁r, Jk (a− b−), the anti-E and anti-Jk^a detected in the alloadsorbed sera must be alloantibodies. Blood for transfusion should be E-negative, Jk^a-negative.

Additional notes on adsorption techniques

1. If the patient has had a transfusion in the past month, an eluate must also be tested because alloantibody may be present on red cells but not in serum/plasma.

2. If the patient’s serum contains a haemolytic antibody, EDTA should be added to prevent the uptake of complement and subsequent lysis of the cells used for adsorption. Add 1 volume of neutral EDTA (potassium salt) (see p. 620) to 9 volumes of serum. More commonly, a plasma sample is used.

3. It is often useful to alloadsorb both serum and eluate to differentiate between autoantibodies and alloantibodies, particularly if the autoantibody is the mimicking type described by Issitt.¹⁸

4. If the autoantibody does not react with papainized cells, do not papainize the cells for adsorption.

Elution of Antibodies from Red Cells

The selection of any elution technique is often based on personal choice and the availability of the necessary reagents and equipment. However, heat elution techniques are best used for the elution of primary cold reactive (IgM) antibodies such as anti-A, anti-N, anti-M and anti-I and IgG anti-A and anti-B antibodies associated with ABO hemolytic disease of the newborn (HDN). The Lui freeze and thaw technique (see below) may also be used for investigation of ABO HDN. Commercially prepared kits that alter the pH of the red blood cells are equally effective and circumvent the hazards of using organic solvents. Refer to the manufacturer’s instructions for details. Methods for heat elution and Lui’s elution techniques are given below. Commercial kits are now widely available.

Notes

1. A large volume of red blood cells is required to obtain enough eluate for testing.

2. The red blood cells must be washed at least six times and the last wash must be kept for testing to ensure removal of all free antibody.

3. Depending on the elution technique used, the prepared eluate may be frozen if testing is not possible immediately after preparation.

Heat Elution

Mix equal volumes of washed packed cells and saline or 6% bovine serum albumin (BSA). Incubate at 56°C for 5 min. Agitate periodically. Centrifuge to pack the red cells. Remove the supernatant (the eluate), which may be haemoglobin stained. Test the eluate ⁴⁸ by appropriate techniques in parallel with the last wash from the red cells.

Freeze and Thaw Elution (Lui)

Mix 0.5 ml of washed packed red cells with 3 drops of PBS or AB serum. Stopper the tube and rotate to coat the glass surface with red cells. Place at –20°C for 10 min. Thaw the red cells rapidly in a 37°C waterbath. Remove the stopper and centrifuge to sediment red cell stroma. Remove the supernatant and test in parallel with the last wash from the red cells by appropriate techniques.⁴⁹

Screening Eluates

The eluate and the saline of the last wash (control) are first screened against two or three samples of washed normal group O cells to see if they contain any antibodies using the IAT. If anti-A or anti-B is suspected, include A and B cells. To 4–6 drops of eluate and control, add 2 drops of a 2% suspension of normal ionic strength saline (NISS).

Incubate for 1–1.5 hours at 37°C. Wash four times and, using optimal dilutions of anti-IgG, carry out the IAT by the tube method.

If the control preparation (the supernatant saline from the last washing) gives positive reactions, the possibility that an eluate contains serum antibody has to be considered.

Determination of the Specificity of Warm Autoantibodies in Eluates and Sera

When tested against a phenotyped panel, about two-thirds of autoantibodies appear to have Rh specificity and in about half of these cases specificity against a particular antigen can be demonstrated.^5,^15,¹⁸ Within the Rh system, anti-e-like is the most common specificity. D − − and Rh_null cells are an advantage.

The other one-third of autoantibodies may show specificity against other, very high incidence antigens (e.g. Wrb and Ena) and rarely other blood-group specificities are involved. It is essential to differentiate between autoantibodies and alloantibodies, especially if transfusion is being considered. The presence of alloantibodies in addition to autoantibodies is suggested by any discrepancy between the serum and eluate results.

As already mentioned, the presence of alloantibodies in a serum complicates the determination of the specificity of an autoantibody and it can be argued that it would be better to test only the eluted autoantibody and to leave the serum strictly alone. However, only a small volume of an eluate may be available, especially in patients who are anaemic and it is generally wise to test both serum and eluate. The procedure is the same for both.

Titration of Warm Antibodies in Eluates or Sera

The methods used are those described in Chapter 21. The exact technique chosen, and the red cells used, should be those that have given the clearest results in the screening tests. Titration of the eluate can be useful in the presence of a panreacting autoantibody to exclude an underlying alloantibody.

In investigating cold autoantibodies, the following tests may sometimes provide clinically useful information.

Determination of the Specificity of Cold Autoantibodies

High-titre cold autoantibodies have a well-defined blood-group specificity, which is very often within the I/i system.^18,^50,⁵¹ Because the I antigen is poorly developed in cord-blood red cells, whereas the i antigen is well developed, group O cord blood red cells should be included in the panel used to test for I/i specificity. Adult cells almost always have the I antigen well expressed, but the strength of the antigen varies and it is of considerable advantage to have available adult cells known to possess strong I antigen. (The rare adult i cells, if available, can also be used.)

Titration of Cold Antibodies

If the screening test is positive for cold auto-agglutinins, titrate as follows.

Prepare doubling dilutions of the serum in saline ranging from 1 in 1 to 1 in 512 and add 1 drop of each serum dilution into three series of (12 × 75 mm) tubes so that three replicate titrations can be made. Add 1 drop of a 2% suspension of pooled saline-washed adult group O (I) cells to the first row, 1 drop of cord-blood group O (i) cells to the second row and 1 drop of the patient’s own cells to the third row. Mix and leave for several hours at 4°C. Before reading, place pipettes and a tray of slides at 4°C. Read macroscopically at room temperature using chilled slides.

Normal range

Using sera from normal White adults and normal adult I red cells, the cold-agglutinin titre at 4°C is 1 to 32; and with cord-blood (i) cells the titre is 0 to 8. In chronic CHAD, the end-point may not have been reached at a dilution of 1 in 512; if that is the case, further dilutions should be prepared and tested.

If a cold agglutinin is present at a raised titre, the presence of a cold alloantibody has to be excluded. In this case, the patient’s own red cells will be found to react much less strongly than do normal adult I red cells. It should be noted that in CHAD the patient’s cells commonly react less strongly than do normal adult I cells (Table 13.8).

Table 13.8 Agglutination titres using various types of cold autoantibodies and normal adult and normal cord red cells, the patient’s red cells and enzyme-treated (papainized) normal adult red cells

Cold Agglutinin Titration Patterns

The presence of high-titre cold agglutinins in a patient’s serum will be indicated by the screening procedure described earlier. To demonstrate that the agglutinins are autoantibodies, it is necessary to show that the patient’s own cells are also agglutinated. The titre using the patient’s cells is usually less than that of control normal adult red cells (one-half or one-quarter) (Table 13.8).

In CHAD, whether ‘idiopathic’ or secondary to mycoplasma pneumonia or lymphoma, the autoantibodies usually have anti-I specificity (Patient A.G. in Table 13.8).

In rare cases of haemolytic anaemia associated with infectious mononucleosis, an autoantibody of anti-i specificity has been demonstrated (Patient F.B. in Table 13.8) and this specificity, too, has been found in certain patients with lymphoma. Rarely, in CHAD, the antibody has been shown to have anti-Pr or anti-M specificity: if enzyme-treated red cells are used, then in either type of case the antigen is destroyed by enzyme treatment (Patient A.R. in Table 13.8).

Determination of the Thermal Range of Cold Agglutinins

From a series of master doubling dilutions of serum in saline, place 1 drop of serum or serum dilution into three rows of (12 × 75 mm) tubes. Set them up at 30°C and at room temperature (20–25°C); to each tube add 1 drop of a 2% saline suspension of the following cells:

1. Pooled normal adult group O (I) red cells

2. Pooled cord-blood group O (i) red cells

3. Patient’s red cells.

Titration should also be carried out at 37°C, if there had been agglutination at this temperature in the screening tests. After incubation at the appropriate temperature for 1 h, determine the presence or absence of agglutination macroscopically over a light.

Detection and Titration of the Donath–Landsteiner Antibody

The D–L antibody of PCH differs from the high-titre cold antibodies referred to previously in that it is an IgG antibody and has a quite different specificity. It is also far more lytic to normal cells in relation to its titre than are anti-I or anti-i. The lysis titre of a D–L antibody may be the same or greater than its agglutination titre. Almost maximal lysis develops in unacidified serum.

Direct Donath–Landsteiner Test

Collect two samples of venous blood into tubes containing no anticoagulant, previously warmed at 37°C. Incubate the first sample at 37°C for 1.5 h. Put the second sample in a beaker packed with ice and allow to stand for 1 h, then place this tube at 37°C for a further 20 min. Centrifuge both tubes at 37°C and examine the supernatant serum for lysis. A positive test is indicated by lysis in the sample that had been chilled. If positive, investigate the antibody specificity (as described later). If negative, proceed to an indirect Donath–Landsteiner test.

A false-negative direct Donath–Landsteiner test result is not uncommon for several reasons:

1. Low antibody level

2. Low complement level (complement is consumed during the haemolytic process)

3. Presence of C3dg on the patient’s red cells (does not lead to complement-mediated haemolysis).

This can be overcome by performing an indirect Donath–Landsteiner test.

Indirect Donath–Landsteiner Test

Serum obtained from the patient’s blood that has been allowed to clot at 37°C is used for this test. Add 1 volume of a 50% suspension of washed normal group O, P-positive red cells to 9 volumes of the patient’s unacidified serum in a tube. Chill the suspension in crushed ice at 0°C for 1 h, then place the tube at 37°C for 30 min. Centrifuge at 37°C and examine for lysis. Three controls should be set up at the same time:

1. A duplicate of the test cell–serum suspension but kept strictly at 37°C for the duration of the test.

2. A duplicate of the test cell–serum suspension, except that an equal volume of ABO-compatible fresh normal serum is first added to the patient’s serum as a source of complement. One volume of the 50% cell suspension is added and the suspension is chilled and subsequently warmed in the same way as the test suspension. (This control excludes false-negative results owing to the patient’s serum being deficient in complement.)

3. A duplicate of the test cell–serum suspension, except that fresh normal serum is used in place of the patient’s serum. This control also is chilled and subsequently warmed.

A positive test will be indicated by lysis in the test suspension and in control No. 2. If ABO-compatible pp cells are available, they should be used in a duplicate set of tubes. No lysis will develop, confirming the P specificity of the antibody.

A false-negative indirect Donath–Landsteiner test can occur. This is a result of the presence of globoside in the serum added as a source of complement. Globoside is the most abundant red cell membrane glycolipid and is present in the serum of all P+ individuals. Addition of ABO-compatible fresh serum as a source of complement could result in cross-reacting with anti-P and this can lead to a false-negative indirect Donath–Landsteiner test. Therefore the indirect Donath–Landsteiner ⁵² test can be modified into two stages.

Two-Stage Indirect Donath–Landsteiner Test

ABO-compatible fresh serum is only added to the red cell–serum suspension after the initial 1-h incubation at 0°C. This prevents antibody inhibition during the cold phase and allows maximum sensitization of the red cells.⁹

Another possible cause of a false-negative indirect Donath–Landsteiner test result is a low antibody level. Papain treatment of the red cells will expose P antigen and can enhance the sensitization.⁵²

Titration of a Donath–Landsteiner Antibody

Prepare doubling or fourfold dilutions of the patient’s serum in fresh normal human serum. To each tube, add a one-tenth volume of a 50% suspension of washed group O, P-positive red cells and immerse each of the tubes in crushed ice at 0°C. After 1 h, place in a 37°C incubator for a further 30 min. Then centrifuge and inspect for lysis.

Detection of a Donath–Landsteiner Antibody by the Indirect Antiglobulin Test

Because the D–L antibody is an IgG antibody, it can be detected by the IAT using an anti-IgG serum if the cells that have been exposed to the antibody in the cold are washed in cold (4°C) saline. At this temperature, the antibody will not be eluted during washing. It should be noted, however, that exposing normal red cells at 4°C to many fresh normal sera results in a positive IAT with broad-spectrum antiglobulin sera because of the adsorption of incomplete anti-H (a normally occurring cold antibody) on to the red cells. At a low temperature, complement is bound too, and it is its adsorption that gives rise to the positive tests with broad-spectrum sera. The adsorption of complement can be prevented by adding a chelating agent, such as EDTA, to the serum.

Method

Add a one-tenth volume of EDTA, buffered to pH 7.0 (see p. 620) to the patient’s serum. Prepare doubling dilutions in saline from 1 in 1 to 1 in 28.

Add 1 volume (drop) of a 50% suspension of group O, P-positive red cells to 10 volumes (drops) of each dilution. Mix and chill at 4°C (preferably in a cold room).

After 1 h, wash the red cells four times in a large volume of cold (4°C) saline. Then carry out an antiglobulin test using an anti-IgG reagent, as described on p. 500, but keeping the red cell-antiglobulin serum suspension at 4°C.

As controls, set up a series of tests using a serum known to contain a D–L antibody (if available) and a normal serum, respectively.

This technique is the most sensitive way of detecting, especially in stored sera, the presence of a D–L antibody in an amount insufficient to bring about actual lysis.

Thermal range of Donath–Landsteiner antibody

The highest temperature at which D–L antibodies are usually adsorbed onto red cells is about 18°C. Hence little or no lysis can be expected unless the cell–serum suspension is cooled below this temperature. Chilling in crushed ice results in maximum adsorption of the antibody and leads to the binding of complement, which brings about lysis when the cell suspension is subsequently warmed at 37°C. Hence the ‘cold-warm’ biphasic procedure necessary for lysis to be demonstrated with a typical D–L antibody.

Specificity of the Donath–Landsteiner antibody

The D–L antibody appears to have a well-defined specificity within the GLOB blood-group system: namely, anti-P. However, in practice, almost all samples of red cells are acted upon because the cells that will not react (P^k and pp) are extremely rare.⁵³ Cord blood cells are lysed to about the same extent as are adult P₁ and P₂ cells.

Treatment of Serum with 2-Mercaptoethanol or Dithiothreitol

Weak solutions of 2-mercaptoethanol (2-ME) or dithiothreitol (DTT) destroy the inter-chain sulphydryl bonds of gamma globulins. IgM antibodies treated in this way lose their ability to agglutinate red cells while IgG antibodies do not.¹⁸ IgA antibodies may or may not be inhibited depending upon whether or not they are made up of polymers of IgA. Since almost all autoantibodies are either IgM or IgG, treatment of serum or an eluate with 2-ME or DTT gives a reliable indication of the Ig class of autoantibody under investigation.^18,⁵⁴

Method

2-Mercaptoethanol

To 1 volume of undiluted serum add 1 volume of 0.1 mol/l 2-ME in phosphate buffer, pH 7.2 (see p. 622).

As a control, add a volume of the serum to the phosphate buffer alone. Incubate both at 37°C for 2 h. Then titrate the treated serum and its control with the appropriate red cells.

If IgG antibody is present, the antibody titration in the control serum will be the same as that of the treated serum. However, if the antibody is IgM, the treated serum will fail to agglutinate the test cells or will agglutinate them to a much lower titre compared with the control untreated serum.

The control must remain active to show that the absence of agglutination is the result of reduction of IgM antibody and not dilution.

Dithiothreitol

0.01 mol/l DTT can be used in place of 0.1 mol/l 2-ME in the previous method.

Drug-Induced Haemolytic Anaemias of Immunological Origin

As already mentioned, acquired haemolytic anaemias may develop as the result of immunological reactions consequent on the administration of certain drugs.^15,⁵⁵^–⁵⁷ Clinically, they often closely mimic AIHA of ‘idiopathic’ origin and for this reason a careful enquiry into the taking of drugs is a necessary part of the interrogation of any patient suspected of having an acquired haemolytic anaemia.

Two immunological mechanisms leading to a drug-induced haemolytic anaemia are recognized. These mechanisms can be referred to as ‘drug-dependent immune’ and ‘drug-induced autoimmune’. Both types of antibody may be present in some patients.^58,⁵⁹ In a unifying concept, the target orientation of these antibodies covers a spectrum in which the primary immune response is initiated by an interaction between the drug or its metabolites and a component of the blood cell membrane to create a neoantigen.⁶⁰ Drug-dependent antibodies bind to both the drug and the cell membrane but not to either separately. If the drug is withdrawn, the immune reaction subsides. It has been postulated that in the case of the autoantibodies, the greater part of the neoantigen is sufficiently similar to the normal cell membrane to allow binding without the drug being present. Similar mechanisms have been described for drug-induced immune thrombocytopenia and neutropenia of immunological origin (see p. 508).

In drug-dependent immune haemolytic anaemia, the drug is required in the in vitro system for the antibodies to be detected. The red cells become damaged by one of two mechanisms:

1. Complement lysis. A typical history is for haemolysis, which may be severe and intravascular, to follow the readministration of a drug with which the patient has previously been treated and for the haemolysis to subside when the offending drug is withdrawn. The DAT is likely to become strongly positive during the haemolytic phase, the patient’s red cells being agglutinated by anti-C′ and sometimes by anti-IgG.

Drugs that have been shown to cause haemolysis by the previously explained mechanism include quinine, quinidine and rifampicin, chlorpropamide, hydrochlorothiazide, nomifensine, phenacetin, salicylazosulphapyridine, the sodium salt of p-aminosalicylic acid and stibophen. Petz and Branch listed 25 drugs reported to have brought about haemolysis by this mechanism.⁵⁶

2. Extravascular haemolysis. This is brought about by IgG antibodies that usually do not activate complement or if they do, not beyond C3. The DAT will be positive with anti-IgG and sometimes also with anti-C′.

The haemolytic anaemia associated with prolonged high-dose penicillin therapy is caused by the previously mentioned mechanism and other penicillin derivatives, as well as cephalosporins and tetracycline, may cause haemolysis in a similar fashion. Haemolysis ceases when the offending drug has been withdrawn.

Cephalosporins, in addition to causing the formation of specific antibodies, may alter the red cell surface so as to cause non-specific adherence of complement and immunoglobulins. This may lead to a positive DAT but is seldom associated with increased haemolysis, although where it occurs it can be very severe.

Drug-Induced Autoimmune Haemolytic Anaemias

In the case of drug-induced autoimmune haemolytic anaemias, the antibody reacts with the red cell in the absence of the drug (these are sometimes referred to as ‘drug-independent antibodies’). The anti-red cell autoantibodies seem to be serologically identical to those of ‘idiopathic’ warm-type AIHA. When the drug was widely used, the great majority of cases followed the use of the antihypertension drug α-methyldopa. The red cells are coated with IgG and the serum contains autoantibodies that characteristically have Rh specificity.

Other drugs that have been reported to act in a similar fashion to α-methyldopa include L-dopa, chlordiazepoxide, mefenamic acid, flufenamic acid and indometacin.⁵

Typical serological features of the different types of drug-induced haemolytic anaemia of immunological origin are summarized in Table 13.9.

Table 13.9 Serological features of the different types of drug-induced haemolytic anaemia of immunological origin

Detection of Antipenicillin Antibodies

The characteristic features of penicillin-induced haemolytic anaemia are as follows:

1. Haemolysis occurs only in patients receiving large doses of a penicillin for long periods (e.g. weeks).

2. The DAT is strongly positive with anti-IgG reagents.

3. The patient’s serum and antibody eluted from the patient’s red cells react only with penicillin-treated red cells – they do not react with normal untreated red cells.

Reagents

Barbitone buffer. 0.14 mol/l, pH 9.5 (see p. 621).

Penicillin solution. 0.4 g of penicillin G dissolved in 6 ml of barbitone buffer.

Penicillin-Coated Normal Red Cells

Wash group O reagent red cells three times in saline and make an approximately 15% suspension in saline to which a one-tenth volume of barbitone buffer has been added. Add 2 ml of the red cell suspension to 6 ml of penicillin solution and incubate at 37°C for 1 h. Then wash four times in saline and make 2% red cell suspensions in saline (for tube tests).

Control normal red cells

Control normal red cells should be treated in exactly the same way as the penicillin-coated red cells except that the 6 ml of penicillin solution is replaced by 6 ml of barbitone buffer.

Method

Antipenicillin antibodies can be detected by the IAT in the usual way using the penicillin-coated red cells in place of normal unmodified cells. However, three extra controls are necessary.

1. Red cells that have not been exposed to penicillin should be added to the patient’s serum.

2. Penicillin-treated red cells should be added to two normal sera known not to contain antipenicillin antibodies (negative controls).

3. Penicillin-treated red cells should be added to a serum (if one is available) known to contain antipenicillin antibodies (positive control).

Cephalosporin can be used in a similar way to sensitize red cells. Control (item 2) is particularly important when drugs such as cephalosporins are used because overexposure in vitro to these drugs can lead to positive results with normal sera.

Note

Some drugs do not dissolve easily; incubation at 37°C, crushing tablets with a pestle and mortar and vigorous shaking of the solution may help.

High-titre IgG antipenicillin antibodies often cause direct agglutination of penicillin-treated red cells in low dilutions of serum. The antibodies can be differentiated from IgM agglutinating antibodies by treatment with 2-ME or DTT (see p. 289).

Detection of Antibodies against Drugs Other than Penicillin

In a patient with an immune haemolytic anaemia whose serum and red cell eluate does not react with normal red cells and who is receiving a drug or drugs other than penicillin or a penicillin derivative, antibodies that react with red cells only in the presence of the suspect drugs or drugs should be looked for in the following way.

The patient’s serum and red cell eluates should be tested with normal and enzyme-treated group O red cells, carrying out the tests with and without the drug that the patient is receiving. The approach is essentially empirical. A saturated solution of the drug or its metabolite should be prepared in saline and the pH should be adjusted to 6.5–7.0.

Set up six tubes containing the patient’s serum and the drug solution in the proportions shown in Table 13.10 and add 1 drop of a 50% saline suspension of group O cells to each tube. Incubate at 37°C for 1 h, then examine for agglutination and lysis. Wash the red cells four times in saline, and carry out an IAT using anti-IgG and anti-C′ separately.

Table 13.10 Investigation of a suspected drug-induced immune haemolytic anaemia

Interpretation

Tubes 1 and 2 test the patient’s serum (? drug-dependent antibody) and normal red cells in the presence of the drug (Tube 1) and without the drug (Tube 2). Tubes 3 and 4 test the effect of added complement on the previous reactions. Tubes 5 and 6 without the patient’s serum act as controls for tubes 3 and 4.

Oxidant-induced haemolytic anaemia

Oxidant-induced haemolytic anaemia should be suspected when the blood film of a patient exposed to an oxidant drug or chemical shows irregularly contracted cells. A Heinz-body test (see p. 336) is confirmatory. The oxidant may also cause methaemoglobinaemia or sulphaemoglobinaemia, both of which can be confirmed by spectroscopy (see p. 240) or co-oximetry. The differential diagnosis of haemolysis induced by an exogenous oxidant includes other causes of haemolysis with irregularly contracted cells (e.g. Zieve’s syndrome), G6PD deficiency and the presence of an unstable haemoglobin. In Zieve’s syndrome (haemolysis associated with alcohol excess, fatty liver and hyper-lipidaemia), the plasma may be visibly lipaemic; if this syndrome is suspected, further investigations should include liver function test and serum lipid measurements.

Microangiopathic and mechanical haemolytic anaemias

Microangiopathic or mechanical haemolytic anaemia should be suspected when a blood film shows schistocytes. Examination of the blood film is, in fact, the most important laboratory procedure in making this diagnosis, although some automated blood cell counters will also detect the presence of red cell fragments. Because haemolysis is intravascular, useful confirmatory tests include serum haptoglobin estimation (see p. 233) and, when the condition is chronic, a Perls’ stain of urinary sediment to detect the presence of haemosiderin (see p. 236). Because a microangiopathic haemolytic anaemia is often part of a more generalized syndrome resulting from microvascular damage or fibrin deposition, other tests are also indicated in unexplained cases. They include tests of renal function, a platelet count and a coagulation screen including tests for D-dimer or fibrin degradation products (see p. 440). Tests for verotoxin-secreting E. coli are indicated in cases of microangiopathic haemolytic anaemia with renal failure. If available, quantification of von Willebrand factor-cleaving protease (ADAMTS13) is indicated in suspected thrombotic thrombocytopenic purpura.

Paroxysmal nocturnal haemoglobinuria

Paroxysmal nocturnal haemoglobinuria (PNH) is an acquired clonal disorder of haemopoiesis, in which the patient’s red cells are abnormally sensitive to lysis by normal constituents of plasma. In its classical form, it is characterized by haemoglobinuria during sleep (nocturnal haemoglobinuria), jaundice and haemosiderinuria. Not uncommonly, however, PNH presents as an obscure anaemia without obvious evidence of intravascular haemolysis or it develops in a patient suffering from aplastic anaemia or more rarely from primary myelofibrosis or chronic myeloid leukaemia.^61,⁶²

PNH red cells are unusually susceptible to lysis by complement.^63,⁶⁴ This can be demonstrated in vitro by a variety of tests (e.g. the acidified-serum [Ham],⁶⁵ sucrose,⁶⁶ thrombin,⁶⁷ cold-antibody lysis,⁶⁸ inulin⁶⁹ and cobra-venom⁷⁰ tests). In the acidified serum, inulin and cobra-venom tests, complement is activated via the alternative pathway, whereas in the cold-antibody test and probably in the thrombin test, complement is activated by the classical sequence initiated through antigen–antibody interaction. In the sucrose lysis test, a low ionic strength is thought to lead to the binding of IgG molecules non-specifically to the cell membrane and to the subsequent activation of complement via the classical sequence. In addition, the alternative pathway appears to be activated.⁷¹ In each test, PNH cells undergo lysis because of their greatly increased sensitivity to lysis by complement.

Minor degrees of lysis may be observed in the cold-antibody lysis and sucrose tests with the red cells from a variety of dyserythropoietic anaemias (e.g. aplastic anaemia, megaloblastic anaemia and primary myelofibrosis).^72,⁷³ Weak positive results in these tests thus have to be interpreted with care. PNH red cells, however, almost always undergo considerable lysis in these tests.

A characteristic feature of a positive test for PNH is that not all the patient’s cells undergo lysis, even if the conditions of the test are made optimal for lysis (Fig. 13.1). This is because only a proportion of any PNH patient’s red cell population is hypersensitive to lysis by complement. This population varies from patient to patient and there is a direct relationship between the proportion of red cells that can be lysed (in any of the diagnostic tests) and the severity of in vivo haemolysis.

Figure 13.1 Effect of pH on lysis in vitro of paroxysmal nocturnal haemoglobinuria (PNH) red cells by human sera. The red cells of three patients with disease of different severity and two fresh normal sera, one serum being more potent than the other, were used.

The phenomenon of some red cells being sensitive to complement lysis and some being insensitive was studied quantitatively by Rosse and Dacie, who obtained two-component complement sensitivity curves in a series of patients with PNH.⁶⁴ Later, Rosse reported that in some cases three populations of red cells could be demonstrated.^74,⁷⁵

1. Very sensitive (type III) cells, 10–15 times more sensitive than normal cells

2. Cells of medium sensitivity (type II), 3–5 times more sensitive than normal cells

3. Cells of normal sensitivity (type I).

In vivo the proportion of type III cells parallels the severity of the patient’s haemolysis.

PNH is an acquired clonal disorder ⁷⁶ resulting from a somatic mutation occurring in a haemopoietic stem cell. It has been demonstrated that a proportion of granulocytes, platelets and lymphocytes are also part of the PNH clone.^77,⁷⁸ The characteristic feature of cells belonging to the PNH clone is that they are deficient in several cell-membrane–bound proteins including red cell acetylcholineesterase,⁷⁹^–⁸¹ neutrophil alkaline phosphatase,⁸²^–⁸⁴ CD55 (decay accelerating factor or DAF),^85,⁸⁶ homologous restriction factor (HRF),^63,⁸⁷ and CD59 (membrane inhibitor of reactive lysis or MIRL),⁸⁸^–⁹⁰ among others. CD55, CD59 and HRF all have roles in the protection of the cell against complement-mediated attack. CD59 inhibits the formation of the terminal complex of complement and it has been established that the deficiency of CD59 is largely responsible for the complement sensitivity of PNH red cells. PNH type III red cells have a complete deficiency of CD59, whereas PNH type II red cells have only a partial deficiency and it is this difference that accounts for their variable sensitivities to complement.^91,⁹² The analysis of these deficient proteins on PNH cells by flow cytometry, particularly of the red cells and neutrophils, has become a useful research and diagnostic tool but is only applicable in centres with a significant number of patients requiring investigation for PNH. By comparing the proportion of cells with deficient CD59 to the percentage lysis in the Ham test, it has been possible to assess the sensitivity of the Ham test. The standard Ham test is reasonably good at estimating the proportion of PNH red cells as long as they are PNH type III cells and comprise <20% of the total. In cases in which the PNH cells are type II and >20% are present, the standard Ham test significantly underestimates the proportion of PNH red cells. The standard Ham test can be negative when there are <5% PNH type III cells or <20% PNH type II cells. When the Ham test is supplemented with magnesium, to optimize the activation of complement, the percentage lysis gives a more accurate estimation of the proportion of PNH cells (Fig. 13.2).⁹³

Figure 13.2 Comparison of the proportion of CD59-deficient red cells with the lysis in the Ham test. The percentage lysis in the Ham test with added magnesium (•) and without added magnesium (o) is plotted against the proportion of CD59-deficient red cells in the same samples from 25 patients with paroxysmal nocturnal haemoglobinuria (PNH).

(With thanks to P. Hillmen, M. Bessler, D. Roper and L. Luzzatto, unpublished observation.)

Certain chemicals, in particular sulphydryl compounds, can act on normal red cells in vitro so as to increase their complement sensitivity. In this way, PNH-like red cells can be created in the laboratory and can be used as useful reagents.

Acidified-Serum Lysis Test (Ham test)

Principle

The patient’s red cells are exposed at 37°C to the action of normal or the patient’s own serum suitably acidified to the optimum pH for lysis (pH 6.5–7.0) (Table 13.11).

Table 13.11 The acidified-serum lysis test with added magnesium

The patient’s red cells can be obtained from defibrinated, heparinized, oxalated, citrated or EDTA blood and the test can be satisfactorily carried out even on cells that have been stored at 4°C for up to 2–3 weeks in ACD or Alsever’s solution, if kept sterile. The patient’s serum is best obtained by defibrination because in PNH if it is obtained from blood allowed to clot in the ordinary way at 37°C or at room temperature, it will almost certainly be markedly lysed. Normal serum should similarly be obtained by defibrination, although serum derived from blood allowed to clot spontaneously at room temperature or at 37°C can be used. Normal serum known to be strongly lytic to PNH red cells is to be preferred to patient’s serum, the lytic potentiality of which is unknown. However, if the test is positive using normal serum, it is important, particularly if the patient appears not to be suffering from overt intravascular haemolysis, to obtain a positive result using the patient’s serum to exclude hereditary erythroid multinuclearity associated with a positive acidified-serum test (HEMPAS) (see p. 294). The variability between the sera of individuals in their capacity to lyse PNH red cells is shown in Figure 13.1. The activity of a single individual’s serum also varies from time to time,⁹⁴ and it is always important to include in any test, as a positive control, a sample of known PNH cells or artificially created ‘PNH-like’ cells (see p. 296).

The sera should be used within a few hours of collection. Their lytic potency is retained for several months at –70°C, but at 4°C, and even at –20°C, this deteriorates within a few days.

Method

Deliver 0.5 ml samples of fresh normal serum, group AB or ABO-compatible with the patient’s blood, into 3 pairs of 75 × 12 mm tubes. Place two tubes at 56°C for 10–30 min to inactivate complement. Keep the other 2 pairs of tubes at room temperature and add to the serum in two of the tubes one-tenth volumes (0.05 ml) of 0.2 mol/l HCl. Add similar volumes of acid to the inactivated serum samples. Then place all the tubes in a 37°C waterbath.

While the serum samples are being dealt with, wash samples of the patient’s red cells and of control normal red cells (compatible with the normal serum) twice in saline and prepare 50% suspensions in the saline. Then add one-tenth volumes of each of these cell suspensions (0.05 ml) to one of the tubes containing unacidified fresh serum, acidified fresh serum and acidified inactivated serum, respectively. Mix the contents carefully and leave the tubes at 37°C. Centrifuge them after about 1 h.

Add 0.05 ml of each cell suspension to 0.55 ml of water so as to prepare a standard for subsequent quantitative measurement of lysis and retain 0.5 ml of serum for use as a blank. For the measurement of lysis, deliver 0.3 ml volumes of the supernatants of the test and control series of cell–serum suspensions and of the blank serum and of the lysed cell suspension equivalent to 0% and 100% lysis, respectively, into 5 ml of 0.4 ml/l ammonia or Drabkin’s reagent. Measure the lysis in a photoelectric colorimeter using a yellow–green (e.g. Ilford 625) filter or in a spectrometer at a wavelength of 540 nm.

If the test cells are from a patient with PNH, they will undergo definite, although incomplete, lysis in the acidified serum. Much less lysis or even no lysis at all, will be visible in the unacidified serum. No lysis will be brought about by the acidified inactivated serum. The normal control sample of cells should not undergo lysis in any of the three tubes.

In PNH, 10–50% lysis is usually obtained when lysis is measured as liberated haemoglobin. Exceptionally, there may be as much as 80% lysis or as little as 5%.

The red cells of a patient who has had a transfusion will undergo less lysis than they would have before the transfusion because the normal transfused cells do not have increased sensitivity to lysis. In PNH, it is characteristic that a young cell (reticulocyte-rich) population, such as the upper red cell layer obtained by centrifugation, undergoes more lysis than the red cells derived from mixed whole blood.

Acidified-Serum Test Lysis with Additional Magnesium (Modified Ham Test)

Principle

The sensitivity of the Ham test can be improved by the addition of magnesium to the test to enhance the activation of complement.

Method

The method is identical to that for the standard Ham test (see above) with the addition of 10 μl of 250 mmol magnesium chloride (final concentration = 4 mmol) to each tube prior to the incubation (Table 13.11).

Significance of the Acidified-Serum Lysis Test

A positive acidified-serum test, carried out with proper controls, denotes the PNH abnormality and PNH cannot be diagnosed unless the acidified-serum test is positive. The addition of magnesium chloride increases the sensitivity of the acidified-serum test.

When the acidified-serum test is positive, a direct antiglobulin test (see p. 500) should also be carried out. If this is positive, it could be the result of a lytic antibody that has given a false-positive acidified-serum test. This can be confirmed by appropriate serological studies. In such complex cases flow cytometry after reaction of the red cells with anti-CD59 is recommended because it is a more definitive test for PNH (see below).

The only disorder other than PNH that may appear to give a clear-cut positive test result is a rare congenital dyserythropoietic anaemia, congenital dyserythropoietic anaemia type II or HEMPAS.^95,⁹⁶ In contrast to PNH, however, HEMPAS red cells undergo lysis in only a proportion (about 30%) of normal sera; moreover, they do not undergo lysis in the patient’s own acidified serum and the sucrose lysis test is negative. In HEMPAS, the expression of glycosylphosphatidylinositol (GPI)-linked proteins, such as CD55 and CD59, is normal. Lysis in HEMPAS appears to be a result of the presence on the red cells of an unusual antigen, which reacts with a complement-fixing IgM antibody (‘anti-HEMPAS’) present in many, but not in all, normal sera.⁹⁶

Heating at 56°C inactivates the lytic system and, if there is lysis in inactivated serum, the test cannot be considered positive. Markedly spherocytic red cells or effete normal red cells may lyse in acidified serum, probably owing to the lowered pH, and such cells may also lyse in acidified inactivated serum.

PNH red cells are not unduly sensitive to lysis by a lowered pH per se. The addition of the acid adjusts the pH of the serum–cell mixture to the optimum for the activity of the lytic system. As is shown in Figure 13.1, it is possible to construct pH–lysis curves, if different concentrations of acid are used. The optimum pH for lysis is between pH 6.5 and 7.0 (measurements made after the addition of the red cells to the serum).

Sucrose Lysis Test

An iso-osmotic solution of sucrose (92.4 g/l) is required.⁹⁷ This can be stored at 4°C for up to 3 weeks.

For the test, set up two tubes, one containing 0.05 ml of fresh normal group AB- or ABO-compatible serum diluted in 0.85 ml of sucrose solution and the other containing 0.05 ml of serum diluted in 0.85 ml of saline. Add to each tube 0.1 ml of a 50% suspension of washed red cells. After incubation at 37°C for 30 min, centrifuge the tubes and examine for lysis. If lysis is visible in the sucrose-containing tube, measure this in a photoelectric colorimeter or a spectrometer as described earlier, using the tube containing serum diluted in saline as a blank and a tube containing 0.1 ml of the red cells suspension in 0.9 ml of 0.4 ml/l ammonia in place of the sucrose–serum mixture as a standard for 100% lysis.

Interpretation

The sucrose lysis test is based on the fact that red cells absorb complement components from serum at low ionic concentrations.^94,⁹⁸ PNH cells, because of their great sensitivity, undergo lysis but normal red cells do not. The red cells from some patients with leukaemia⁷² or primary myelofibrosis may undergo a small amount of lysis, almost always <10%; in such cases, the acidified-serum test is usually negative and PNH should not be diagnosed. In PNH, lysis is usually between 10% and 80%, but exceptionally may be as little as 5%. Sucrose lysis and acidified-serum lysis of PNH red cells are fairly closely correlated. The sucrose lysis test is usually negative in HEMPAS.

Flow Cytometry Analysis of the GPI-Linked Proteins on Red Cells

Principle

The patient’s red cells are stained with a fluorescein-labelled antibody that is specific for one of several GPI-linked proteins, for example CD55, CD58 (leucocyte function antigen-3) or CD59 (‘protectin’), which are deficient in PNH red cells. The stained cells are then analysed with a flow cytometer. In some laboratories where facilities are available, flow cytometry has replaced the Ham test as the primary method for the diagnosis of PNH.

The patient’s red cells can be obtained in any of the anticoagulants described for the Ham test. The cells, if taken into ACD, can be stored for 2–3 weeks prior to analysis. Fluorescein-conjugated anti-CD59 (MEM 43, Cymbus Bioscience Ltd, Southampton, UK) gives excellent results when used for red cell analysis. It is important to use the conjugated antibody because staining with unconjugated anti-CD59 followed by a fluorescein-conjugated second layer antibody may result in artefact, probably owing to red cell agglutination. There is no suitable anti-CD55 antibody commercially available at present. Anti-CD58 (BRIC5, Bioproducts Laboratories, UK, used at 20 μg/ml) gives reproducible results for red cell analysis, but the level of CD58 expression on PNH type-II cells is higher than that of many other GPI-linked proteins and thus studying CD58 expression is useful but not ideal.⁹⁹

Method

Chill 1 × 10⁶ cells in 50 μl of PBS on ice with 50 μl of monoclonal antibody for 30 min. Wash twice in PBS + azide (200 mg/l) and then chill with fluorescein-labelled goat antimouse antibody on ice in the dark for 30 min. Wash twice in PBS + azide and then fix with approximately 0.5 ml of 1% formaldehyde in Isoton II (Beckman Coulter). Analysis is performed using a flow cytometer. A negative control antibody should always be used to assess the fluorescence of cells lacking the antigen. The cells from a normal subject should be stained as an additional control to verify that negative cells in the test sample are true PNH cells and not artefactual.

Other Immunological Techniques

The GPI-linked proteins such as CD59 can also be studied by a modification of the gel technology used for blood grouping.

Flow Cytometry Analysis of the GPI-Anchor or GPI-Linked Proteins on Neutrophils

Principle

A proportion of the patient’s neutrophils have been demonstrated to be part of the PNH clone in all patients with PNH. GPI-linked proteins that are suitable for analysis include CD16, CD24, CD55, CD59 and CD67.^100,¹⁰¹ There are available numerous fluorescein-conjugated antibodies to CD16 that are suitable for use in this analysis – for example, fluorescein-conjugated anti-Leu-11a (Beckman Coulter). Similarly, fluorescein-conjugated anti-CD59 (Cymbus Biosciences) and anti-CD55 are of value. FLAER (fluorescein-labelled pro-aerolysin) analysis can also be used, pro-aerolysin binding selectively and with high affinity to the GPI anchor.¹⁰²

Method

The patient’s neutrophils are obtained by collecting blood, anticoagulating with preservative-free heparin (10 μ/ml), and obtaining a buffy coat. The formation of a buffy coat can be accelerated by adding 1 ml of 6% hetastarch in 0.9% sodium chloride (Hespan, DuPont) to 10 ml of blood. Then 1–2 × 10⁶ cells are analysed. It is important that all the subsequent staining and washing are performed at 4°C to minimize non-specific staining. Chill the cells in 50 μl of PBS on ice with 50 μl of monoclonal antibody (MoAb) for 30 min. Wash twice in PBS + 0.1% BSA (PBS + BSA) and then chill with fluorescein-labelled goat antimouse antibody on ice in the dark for 30 min. Wash twice in PBS + BSA and then fix with approximately 0.5 ml of 1% formaldehyde in Isoton II (Beckman Coulter). For conjugated antibodies, a single incubation step only is required, followed by a wash and then fixing prior to analysis. Analysis is performed using a flow cytometer. Appropriate normal controls and negative controls should always be tested in parallel to the patient’s samples.

A negative control antibody should always be used to assess the fluorescence of cells lacking the antigen. The cells from a normal subject should be stained as an additional control to verify that negative cells in the test sample are true PNH cells and not an artefact.

Significance of Flow Cytometric Analysis

The presence of a population of cells with a deficiency of more than one GPI-linked protein is diagnostic of PNH (Fig. 13.3). It is important to analyse more than one protein, because there are extremely rare cases in which an inherited deficiency of one protein has been described (i.e. the Inab phenotype,¹⁰³^–¹⁰⁵ a deficiency of CD55 owing to a defect of the structural gene encoding this protein, and inherited deficiency of CD59¹⁰⁶ due to a defect in the gene encoding CD59). Analysis of the expression of CD59 on erythrocytes allows the identification of PNH type II as well as PNH type III red cells. This is important because, although patients with only PNH type II red cells do not usually suffer from significant haemolysis, they may suffer some of the complications of PNH, such as thrombosis. The analysis of neutrophils for GPI-linked proteins is more difficult than red cell analysis. It is, however, probably more sensitive because the proportion of abnormal neutrophils is usually higher than the proportion of PNH red cells because of the reduced survival of PNH red cells compared to normal cells and because of the effect of transfusions. Thus, flow cytometry applied to neutrophils is a more sensitive method for the diagnosis of PNH than methods relying on the complement sensitivity of PNH red cells.