The various initial steps to be taken in the investigation of a patient suspected of having a haemolytic anaemia are outlined in Chapter 11 and the changes in red cell morphology that may be found in haemolytic anaemias are illustrated in Chapter 5. This chapter describes procedures useful in investigating haemolytic anaemias suspected to result from defects within the red cell membrane or deficiency of enzymes important in red cell metabolism.

The precise identification of an enzyme defect is beyond the scope of most haematology laboratories; it may require the isolation and purification of the enzyme and the determination of its kinetic and structural properties. In a service laboratory, it is sufficient to identify the general nature of the defect, whether it be in the membrane or the metabolic pathways of the red cell. In the case of putative metabolic defects, an attempt should be made, where possible, to pinpoint the enzyme involved. The first part of this chapter describes screening tests for spherocytosis, including hereditary spherocytosis (HS) and for glucose-6-phosphate dehydrogenase (G6PD) deficiency. The later sections of the chapter describe specific enzyme assays and the measurement of 2,3-diphosphoglycerate (2,3-DPG) and reduced glutathione (GSH).

Most of the enzyme assays have been standardized by the International Council for Standardization in Haematology (ICSH). Commercial kits are also available for some quantitative assays and screening tests. These are noted in the relevant sections.

Investigation of membrane defects

The osmotic fragility test gives an indication of the surface area/volume ratio of erythrocytes. Its greatest usefulness is in the diagnosis of hereditary spherocytosis. The test may also be used in screening for thalassaemia. Red cells that are spherocytic, for whatever cause, take up less water in a hypotonic solution before rupturing than do normal red cells.

Other tests that demonstrate red cell membrane defects include glycerol lysis time, cryohaemolysis, autohaemolysis and, more specifically, membrane protein analysis.

Procedures to assess red cell flexibility (rigidity) using polycarbonate membrane filtration,¹ and red cell deformability measurements on specialized equipment such as the Laser-assisted Optical Rotational Cell Analyser (LORCA)² have been described elsewhere.

Osmotic fragility as measured by lysis in hypotonic saline

Principle

The method to be described is based on that of Parpart et al.³ Small volumes of blood are mixed with a large excess of buffered saline solutions of varying concentration. The fraction of red cells lysed at each saline concentration is determined colorimetrically. The test is normally carried out at room temperature (15–25°C).

Reagents

Prepare a stock solution of buffered sodium chloride, osmotically equivalent to 100 g/l (1.71 mol/l) NaCl, as follows: dissolve NaCl, 90 g; Na₂HPO₄, 13.65 g (or Na₂HPO₄2H₂O,17.115 g); and NaH₂PO₄.2H₂O, 2.34 g in water. Adjust the final volume to 1 litre. This solution will keep for months at 4°C in a well-stoppered bottle. Salt crystals may form on storage and must be thoroughly redissolved before use.

In preparing hypotonic solutions for use, it is convenient to make first a 10 g/l solution from the 100 g/l NaCl stock solution by dilution with water. Dilutions equivalent to 9.0, 7.5, 6.5, 6.0, 5.5, 5.0, 4.0, 3.5, 3.0, 2.0 and 1.0 g/l are convenient concentrations. Intermediate concentrations such as 4.75 and 5.25 g/l are useful in critical work and an additional 12.0 g/l dilution should be used for incubated samples.

It is convenient to make up 50 ml of each dilution. The solutions keep well at 4°C if sterile, but should be inspected for moulds before use and discarded if moulds develop.

Method

Heparinized venous blood or defibrinated blood may be used; oxalated or citrated blood is not suitable because of the additional salts added to it. The test should be carried out within 2 h of collection with blood stored at room temperature or within 6 h if the blood has been kept at 4°C.

1. Deliver 5.0 ml of each of the 11 saline solutions into 12 × 75 mm test tubes. Add 5.0 ml of water to the 12th tube.

2. Add to each tube 50 μl of well-mixed blood and mix immediately by inverting the tubes several times, avoiding foam.

3. Leave the suspensions for 30 min at room temperature. Mix again and then centrifuge for 5 min at 1200 g.

4. Remove the supernatants and estimate the amount of lysis in each using a spectrometer at a wavelength setting of 540 nm or a photoelectric colorimeter provided with a yellow-green (e.g. Ilford 625) filter. Use as a blank the supernatant from tube 1 (osmotically equivalent to 9 g/l NaCl).

5. Assign a value of 100% lysis to the reading with the supernatant of tube 12 (water) and express the readings from the other tubes as a percentage of the value of tube 12. Plot the results against the NaCl concentration (Fig. 12.1).

Figure 12.1 Osmotic fragility curves. Osmotic fragility curves of patients suffering from the following: sickle cell anaemia, β thalassaemia major, hereditary spherocytosis and ‘idiopathic’ warm autoimmune haemolytic anaemia. The normal range is indicated by the unbroken lines.

Notes

1. The measurement of osmotic fragility is a simple procedure that requires a minimum of equipment. It will yield gratifying results if carried out carefully.

2. The blood must be delivered into the 12 tubes with great care. The critical point is not that the amount be exactly 50 μl, but rather that the amount added to each tube must be the same. Two methods are recommended:

a. Using an automatic pipette, after aspirating the blood gently, wipe the outside with tissue paper, taking care not to suck out any blood from the inside of the tip by capillary action. The blood is then delivered into the saline solution and the pipette is rinsed in and out several times until no blood is visible inside its tip.

The tip has to be changed before moving on to the next tube. This procedure takes time and may result in an increased exposure for the first few tubes. It is therefore advisable to start the timing only after the addition of the sample to the first tube.

b. Using a Pasteur pipette with a perfectly flat end, 1 mm in diameter, suck up about 1 ml of blood, avoiding any bubbles and wipe the outside of the pipette. With the pipette held vertically above tube 1, deliver a single drop (about 50 μl) without the blood touching the wall of the tube. Then deliver single drops into the remaining 11 tubes.

Method b appears to be primitive, but with practice it is perfectly satisfactory; it is also more economical and much faster than Method a. With either method, the best way to test its accuracy is to do a preliminary test by delivering the blood into several tubes all containing the same saline solution (e.g. either 3.0 or 1.0 g/l). The readings with the supernatants should be all within 5% of each other.

3. If the amount of blood available is limited (e.g. from babies) and the spectrometer takes 1 ml cuvettes, the volumes can be scaled down to 1 ml of saline solution and 10 μl of blood. However, to deliver 10 μl of blood reproducibly is not easy. With Method b, a Pasteur pipette or capillary pipette with a much smaller diameter, calibrated to give 10 μl drops of blood, would have to be used. It is then more difficult to maintain accuracy. Method a may be preferable in this case.

4. With the method using 50 μl of blood and with non-anaemic blood, the reading for 100% lysis will be about 0.7. With a modern spectrometer, any figure between 0.5 and 1.5 is acceptable. If the value is lower than 0.5, the test should be repeated using more blood or less saline (the reverse if the reading is higher than 1.5). With photoelectric colorimeters, values higher than 0.5 are often not very accurate.

5. When transferring the supernatant from a tube to the spectrometer cuvette, care has to be taken not to disturb the pellet. If it is well packed, the supernatant can simply be poured from the tube into the cuvette; with a spectrometer provided with an automatic suction device, this is usually satisfactory. Alternatively, a plastic Pasteur pipette should be used.

6. Even when a normal range has been established, it is essential always to run a normal control sample along with that of the patients to be tested to check, for example, the saline solutions.

The sigmoid shape of the normal osmotic fragility curve indicates that normal red cells vary in their resistance to hypotonic solutions. Indeed, this resistance varies gradually (osmotically) as a function of red cell age, with the youngest cells being the most resistant and the oldest cells being the most fragile. The reason for this is that old cells have a higher sodium content and a decreased capacity to pump out sodium.

Osmotic Fragility after Incubating the Blood at 37°C for 24 Hours

Method

Defibrinated blood should be used, care being taken to ensure that sterility is maintained.

Incubate 1 ml or 2 ml volumes of blood in sterile 5 ml bottles. It is advisable to set up the samples in duplicate in case one has become infected, as indicated by gross lysis and change in colour.

After 24 h, if no infection is evident, pool the contents of the duplicate bottles after thoroughly mixing the sedimented red cells in the overlying serum and estimate the fragility as previously described.

Because the fragility may be markedly increased (Fig. 12.2), set up additional hypotonic solutions containing 7.0 g/l and 8.0 g/l NaCl. In addition, use a solution equivalent to 12.0 g/l NaCl because sometimes, as in HS, lysis may take place in 9.0 g/l NaCl. In this case, use the supernatant of the tube containing 12.0 g/l NaCl as the blank in the colorimetric estimation.

Figure 12.2 Osmotic fragility curves before and after incubating blood at 37°C for 24 h. Results are from patients suffering from the following: hereditary spherocytosis, pyruvate kinase deficiency and hereditary non-spherocytic haemolytic anaemia of undiagnosed type. The normal range is indicated by the unbroken lines.

The incubation fragility test is conveniently combined with the estimation of the amount of spontaneous autohaemolysis (p. 252).

Factors Affecting Osmotic Fragility Tests

In carrying out osmotic fragility tests by any method, three variables capable of markedly affecting the results must be controlled, quite apart from the accuracy with which the saline solutions have been made up. These are as follows:

1. The relative volumes of blood and saline

2. The final pH of the blood in saline suspension

3. The temperature at which the tests are carried out.

A proportion of 1 volume of blood to 100 volumes of saline is chosen because the concentration of blood is so small that the effect of the plasma on the final tonicity of the suspension is negligible. When weak suspensions of blood in saline are used, it is necessary to control the pH of the hypotonic solutions and it is for this reason that phosphate buffer is added to the saline. Even so, small differences will be found between the fragility of venous blood and maximally aerated (i.e. oxygenated) blood. For the most accurate results, it is recommended that the blood should be mixed until bright red. Finally, it is ideal for tests to be carried out always at the same temperature, although for most purposes room temperature is sufficiently constant.

The extent of the effect of pH and temperature on osmotic fragility was well illustrated in the paper of Parpart et al.³ The effect of pH is more important: a shift of 0.1 of a pH unit is equivalent to altering the saline concentration by 0.1 g/l, the fragility of the red cells being increased by a decrease in pH. An increase in temperature decreases the fragility, an increase of 5°C being equivalent to an increase in saline concentration of about 0.1 g/l.

Lysis is virtually complete at the end of 30 min at 20°C and the hypotonic solutions may be centrifuged at the end of this time.

Further details of the factors that affect and control haemolysis of red cells in hypotonic solutions were given by Murphy.⁴

Recording the Results of Osmotic Fragility Tests

In the past, osmotic fragility most often has been expressed in terms of the highest concentration of saline at which lysis is just detectable (initial lysis or minimum resistance) and the highest concentration of saline in which lysis appears to be complete (complete lysis or maximum resistance). It is, however, useful also to record the concentration of saline causing 50% lysis (i.e. the median corpuscular fragility, MCF) and to inspect the entire fragility curve (Fig. 12.1). The findings in health are summarized in Table 12.1.

Table 12.1 Osmotic fragility in health (at 20°C and pH 7.4)

	Fresh blood (g/l NaCl)	Blood incubated 24 h, 37°C (g/l NaCl)
Initial lysis	5.0	7.0
Complete lysis	3.0	2.0
MCF (50% lysis)	4.0–4.45	4.65–5.9

MCF, median corpuscular fragility.

Alternative methods of recording osmotic fragility

Two simple alternative methods of recording the results quantitatively are available: the data may be plotted on probability paper or increment-haemolysis curves can be drawn. Both methods emphasize heterogeneity of the cell population with respect to osmotic fragility. If the observed amounts of lysis of normal blood are plotted on the probability scale against concentrations of saline, an almost straight line can be drawn through the points; the line is only skewed where lysis is almost complete. This method enables the MCF to be read off with ease.

In disease, tailed curves also skew the line to varying degrees at the other end of the probability plot. To obtain increment-haemolysis curves, the differences in lysis between adjacent tubes are plotted against the corresponding saline concentrations. Definitely bimodal curves may be obtained during recovery from a haemolytic episode.⁵

Interpretation of Results

The osmotic fragility of freshly taken red cells reflects their ability to take up a certain amount of water before lysing. This is determined by their volume-to-surface area ratio. The ability of the normal red cell to withstand hypotonicity results from its biconcave shape, which allows the cell to increase its volume by about 70% before the surface membrane is stretched; once this limit is reached lysis occurs.⁶ Spherocytes have an increased volume-to-surface area ratio; their ability to take in water before stretching the surface membrane is thus more limited than normal and they are therefore particularly susceptible to osmotic lysis. The increase in osmotic fragility is a property of the spheroidal shape of the cell and is independent of the cause of the spherocytosis. Characteristically, osmotic fragility curves from patients with HS who have not been splenectomized show a ‘tail’ of very fragile cells (Fig. 12.3). When plotted on probability paper, the graph indicates two populations of cells: the very fragile and the normal or slightly fragile. After splenectomy the red cells are more homogeneous, the osmotic fragility curve indicating a more continuous spectrum of cells, from fragile to normal.

Figure 12.3 Osmotic fragility curves of three patients suffering from hereditary spherocytosis belonging to the same family (brother, sister and uncle [Le Ay]). The area between the unbroken lines represents the normal range.

Decreased osmotic fragility indicates the presence of unusually flattened red cells (leptocytes) in which the volume-to-surface area ratio is decreased. Such a change occurs in iron deficiency anaemia and thalassaemia in which the red cells with a low mean cell haemoglobin (MCH) and mean cell volume (MCV) are unusually resistant to osmotic lysis (Fig. 12.1). A simple one-tube osmotic fragility is a useful screening test for β thalassaemia and some haemoglobinopathies in countries with a high incidence of these abnormalities (p. 612). Reticulocytes and red cells from patients who have been splenectomized also tend to have a greater amount of membrane compared with normal cells and are osmotically resistant. In liver disease, target cells may be produced by passive accumulation of lipid and these cells, too, are resistant to osmotic lysis.⁷

The osmotic fragility of red cells after incubation for 24 h at 37°C is also a reflection of their volume-to-surface area ratio, but the factors that alter this ratio are more complicated than in fresh red cells. The increased osmotic fragility of normal red cells, which occurs after incubation (Fig. 12.2), is mainly caused by swelling of the cells associated with an accumulation of sodium that exceeds loss of potassium. Such cation exchange is determined by the membrane properties of the red cell, which control the passive flux of ions, and the metabolic competence of the cell, which determines the active pumping of cations against concentration gradients. During incubation for 24 h, the metabolism of the red cell becomes stressed and the pumping mechanisms tend to fail, one factor being a relative lack of glucose in the medium.

The osmotic fragility of red cells that have an abnormal membrane, such as those of HS and hereditary elliptocytosis (HE), increases abnormally after incubation (Fig. 12.2). Similar results occur in hereditary stomatocytosis.⁸ The results with red cells with a glycolytic deficiency, such as those of pyruvate kinase (PK) deficiency, are variable. In severe deficiencies, osmotic fragility may increase substantially (Fig. 12.2), but, in other cases, the fragility may decrease owing to a greater loss of potassium than gain of sodium. In thalassaemia major and minor, osmotic fragility is frequently markedly reduced after incubation, again owing to a marked loss of potassium.⁹ A similar, although usually less marked, change is seen in iron deficiency anaemia.

To summarize, measurement of red cell osmotic fragility provides a useful indication as to whether a patient’s red cells are normal because an abnormal result invariably indicates abnormality. The reverse is, however, not true (i.e. a result that is within the normal range does not mean that the red cells are normal). The findings in some important haemolytic anaemias are summarized in Table 12.2.

Table 12.2 Osmotic fragility in haemolytic anaemias: a summary

Condition	Notes
A. Associated with increased osmotic fragility (OF)
Hereditary spherocytosis (HS)	Entire curve may be ‘shifted to the right’, or most of it may be within the normal range but with a ‘tail’ of fragile cells. Curve within normal range in 10–20% of cases. After incubation for 24 h, abnormalities usually more marked, but still some false-negative results. Splenectomy does not affect MCF but reduces the tail of fragile cells
Hereditary elliptocytosis (HE)	As in HS, but in general changes less marked. Abnormal OF usually correlates with severity of haemolysis (i.e. OF is normal in non-haemolytic HE)
Hereditary stomatocytosis	As in HS with large osmotically fragile cells with low MCHC
Other inherited membrane abnormalities	Results variable; with milder disorders curve more likely to be abnormal after incubation for 24 h
Autoimmune haemolytic anaemia	Tail of fragile cells roughly proportional to number of spherocytes; rest of curve normal (or even left-shifted on account of reticulocytosis)
B. Associated with decreased OF
Thalassaemia	MCF decreased in all forms of thalassaemia, except in some α thalassaemia heterozygotes; usually the entire curve is left-shifted
Enzyme abnormalities	OF usually normal (anaemia originally referred to as hereditary nonspherocytic), but tail of highly resistant cells may be seen on account of high reticulocyte court. After incubation for 24 h, there may be a tail of fragile cells
Hereditary xerocytosis	Increased resistance to osmotic lysis and increased MCHC
Iron deficiency	Curve shifted to left, wholly or partly, depending on proportion of hypochromic red cells

OF, osmotic fragility; HE, hereditary elliptocytosis; HS, hereditary spherocytosis; MCF, median corpuscular fragility; MCHC, mean cell haemoglobin concentration.

Flow cytometric (dye-binding) test

Principle

The osmotic fragility test lacks specificity and sensitivity and fails to detect atypical or mild HS. Moreover, it can be affected by factors unrelated to red cell cytoskeletal defects; for example, positive results may be obtained for red cells from patients who are pregnant or who have immune or other haemolytic anaemias or renal failure. The flow cytometric (dye-binding) test of King and colleagues ¹⁰ measures the fluorescent intensity of intact red cells labelled with eosin-5-maleimide (EMA), which reacts covalently with Lys-430 on the first extracellular loop of Band 3 protein. The N-terminal cytoplasmic domain of Band 3 interacts with ankyrin and protein 4.2, which in turn crosslink with the spectrin-based cytoskeleton and stabilizes the membrane lipid bilayer.¹¹ Deficiency or abnormality of Band 3 may result in decreased fluorescence. This is seen in HS red cells but has also been observed in cases of South-east Asian ovalocytosis, congenital dyserythropoietic anaemia Type II and the stomatocytic variant, cryohydrocytosis. Blood samples in ethylenediaminetetra-acetic acid (EDTA) may be analysed for up to 48 h after collection provided they have been stored in the refrigerator.

Method

Thaw a tube of stock EMA solution in the dark at room temperature and dilute with an equal volume of PBS to obtain a working solution of 0.5 mg/ml. Mix 5 μl of washed packed red cells with 25 μl of EMA working solution in a plastic microfuge tube. Set up control tubes from blood of normal individuals and perform all tests in duplicate. Leave a rack with the tubes in a cupboard in the dark at room temperature for 30 min. Mix and return to the cupboard for a further 30 min.

Then, spin the tubes in a bench-top microfuge for 5–10 s and remove the supernatant dye carefully with a fine-tip pipette.

Wash the labelled red cells three times with 500 μl PBS containing 0.5% BSA. The third wash should be colourless. If it is still pink, suggesting that traces of dye particle remain in the tube, discard the sample and repeat the cell labelling procedure.

Resuspend the packed red cells in 500 μl of the PBS–BSA wash solution. Transfer 100 μl of the cell suspension into a plastic flow cytometer tube and add 1.4 ml of wash solution. Keep the cell suspensions in the dark by wrapping in aluminium foil until use. Set the analyser thresholds (gate) for red blood cells and count each sample for a minimum of 15 000 events. Select the FITC (fluorescein isothiocyanate) channel and record the mean fluorescence intensity (MFI). Compare the test with the mean value of several control samples.

Interpretation of Results

Results should be expressed as a ratio of mean value of the test to control sample. Each laboratory should set the reference range and cut-off values for its own instrument. In our laboratory, a ratio of 0.8 or more is regarded as normal.

Glycerol lysis-time tests

The osmotic fragility test is somewhat cumbersome and requires 2 ml or more of whole blood. It is thus not suitable for use in newborn babies or as a population screening test. In 1974, Gottfried and Robertson ¹² introduced a glycerol lysis-time (GLT) test, a one-tube test, to measure the time taken for 50% haemolysis of a blood sample in a buffered hypotonic saline–glycerol mixture. The original method had greater sensitivity in the osmotic-resistant range, but it also could identify most patients with HS by a shorter GLT₅₀. Better identification of HS blood from normal was obtained by 24-h incubation of samples and by modifying the glycerol reagent.¹³ Zanella et al modified the original test further by decreasing the pH.¹⁴ There is some loss of specificity for HS with the acidified glycerol lysis-time test (AGLT) compared with the original method, but in practice this loss is unimportant.

Acidified Glycerol Lysis-Time Test

Cryohaemolysis Test

Principle

Whereas osmotic fragility may be abnormal in any condition where spherocytes occur, it has been suggested that cryohaemolysis is specific for HS.¹⁵ This appears to result from the fact that the latter is dependent on factors that are related to molecular defects of the red cell membrane rather than to changes in the surface area-to-volume ratio. The test can be carried out on EDTA blood up to 1 day old.

Reagent

Buffered 0.7 mol/l sucrose

23.96 g sucrose in 100 ml of 50 mmol/l phosphate buffer, pH 7.4. This can be stored frozen in 2 ml aliquots in tubes ready for use.

Method ¹⁵

1. Centrifuge the blood and wash the red cells three times with cold (4°C) 9 g/l NaCl. Make a suspension of 50–70% cells in the saline and keep on ice until tested.

2. Prepare 2 ml volumes of reagent, thawing if frozen, and stand for 10 min in a 37°C waterbath to equilibrate.

3. Pipette 50 μl of the cell suspension into each of 2 tubes of the warmed reagent, vortex immediately for a few seconds and then incubate for exactly 10 min at 37°C.

4. Without delay, transfer the tubes to an icebath for another 10 min, vortex for a few seconds and then centrifuge to sediment the remaining cells. Transfer some of the supernatant to a clean tube.

5. Prepare a 100% haemolysate solution by pipetting 50 μl of the original sample into 2 ml of water. Centrifuge and dilute 200 μl of the supernatant in 4 ml of water.

6. Read absorbance at 540 nm of the test and the 100% lysis samples.

Interpretation

Streichman et al.¹⁵ report the range of cryohaemolysis in normal subjects to be 3–15%, whereas in hereditary spherocytosis there is >20% lysis. However, it is recommended that individual laboratories establish their own reference values for the method. We have found that most normal samples give <3% lysis. Increased lysis is not exclusive to hereditary spherocytosis and may be observed in hereditary stomatocytosis.

Autohaemolysis: spontaneous haemolysis developing in blood incubated at 37°c for 48 hours

The autohaemolysis test is useful as an initial screen in suspected cases of haemolytic anaemia. It provides information about the metabolic competence of the red cells and helps to distinguish membrane and enzyme defects if the results of the tests are taken together with other observations such as morphology, inheritance and the presence or absence of associated clinical disorders.¹⁶

Principle

Aliquots of blood are incubated both with and without sterile glucose solution at 37°C for 48 h. After this period, the amount of spontaneous haemolysis is measured colorimetrically.

Method

It is essential to use aseptic techniques in setting up the autohaemolysis test to maintain sterility throughout the incubation period.

Defibrinating Blood

Defibrinate blood, as described on p. 5.

Use sterile defibrinated blood and deliver four 1 ml or 2 ml samples into sterile 5 ml capped bottles. Retain a portion of the original sample; separate and store this as the preincubation serum.

Add to two of the bottles 50 or 100 μl of sterile 100 g/l glucose solution, so as to provide a concentration of glucose in the blood of at least 30 mmol/l. Make sure that the caps of the bottles are tightly closed and place the series of bottles in the incubator at 37°C. A sample from a known normal individual should be run in parallel as a control.

After 24 h, thoroughly mix the content by gentle swirling. After incubating for 48 h, inspect the samples for signs of infection, thoroughly mix again, then from each bottle remove a sample for the estimation of the packed cell volume (PCV) (by the microhaematocrit method) and haemoglobin (Hb) concentration and centrifuge the remainder to obtain the supernatant serum.

Estimate the spontaneous lysis by means of a colorimeter or a spectrometer at 540 nm.

As a rule, it is convenient to make a 1 in 10 dilution of the incubated serum in cyanide-ferricyanide (Drabkin’s) solution (p. 25), unless there is marked haemolysis, when a 1 in 25 or 1 in 50 dilution is more suitable. A corresponding dilution of the preincubation serum is used as a blank and a 1 in 100 or 1 in 200 dilution of the whole blood in Drabkin’s solution indicates the total amount of Hb present and serves as a standard.

Calculate the percentage lysis, allowing for the change in PCV resulting from the incubation as follows:

where R₀ = reading of diluted whole blood; R_t = reading of diluted serum at 48 h; B = reading of blank; PCV_t = packed cell volume at time t; D₀ = dilution of whole blood (e.g. 1 in 200 = 0.005); and D_t = dilution of serum (e.g. 1 in 10 = 0.1).

The reading at time t is multiplied by (1−PCV_t) so as to give the concentration that would be found if the liberated Hb was dissolved in whole blood (i.e. in both plasma and red cell compartments), not in the plasma compartment alone.

Normal Range of Autohaemolysis

Lysis at 48 h: without added glucose, 0.2–2.0%; with added glucose, 0–0.9%.

The results obtained are sensitive to slight differences in technique and each laboratory should use a carefully standardized procedure and establish its own normal range. If the amount of liberated Hb is small, it is more accurate (although more time consuming) to measure lysis by a chemical method rather than by a direct spectrometric method (p. 231). It can also be measured directly by a simple and rapid procedure with a HemoCue Plasma/Low Hb system.¹⁷

Significance of Increased Autohaemolysis

Little or no lysis takes place when normal blood is incubated for 24 h under sterile conditions and the amount present after 48 h is small.¹⁶ If glucose is added so that it is present throughout the incubation, the development of lysis is markedly slowed. The amount of autohaemolysis that occurs after 48 h with and without glucose is determined by the properties of the membrane and the metabolic competence of the red cell. In membrane disorders such as HS, the rate of glucose consumption is increased to compensate for an increased cation leak through the membrane.⁸ During the 48-h incubation, glucose is therefore used up relatively rapidly so that energy production fails more quickly than normal unless glucose is added. This is one factor that contributes to the increased rate of autohaemolysis in HS. Usually, but not always, the addition of glucose to the blood decreases the rate of autohaemolysis in HS. This was referred to as Type 1 autohaemolysis.¹⁶ When the utilization of glucose via the glycolytic pathway is impaired, as in PK deficiency, the rate of autohaemolysis at 48 h is usually increased but glucose fails to correct or may even aggravate lysis (Type 2 autohaemolysis).⁸ Although a similar result may be seen in severe HS (Type B), in the absence of spherocytosis failure of glucose to diminish autohaemolysis is a strong indication of a glycolytic block. Blood from patients with G6PD deficiency or other disorders of the pentose phosphate pathway may undergo a slight increase in autohaemolysis (without additional glucose), which is corrected by the addition of glucose. Commonly, the result is normal, but examination of the incubated blood may show an increase in methaemoglobin (Hi) (discussed later). Not all glycolytic enzyme deficiencies give a Type 2 reaction so that a Type 1 result does not exclude the possibility of such a defect.

In the acquired haemolytic anaemias, the results of the autohaemolysis test are variable and generally not very helpful in diagnosis. In the autoimmune haemolytic anaemias, lysis may be increased in the absence of additional glucose but the effect of added glucose is unpredictable. In paroxysmal nocturnal haemoglobinuria (PNH), the autohaemolysis of aerated defibrinated blood is usually normal.

Autohaemolysis may be increased in haemolytic anaemias caused by oxidant drugs or when there are defects in the reducing power of the red cell. Heinz bodies, Hi or both will be detectable at the end of incubation. Normally, red cells produce <4% Hi after 48 h incubation and Heinz bodies are not seen. Red cells containing an unstable Hb also contain Heinz bodies at the end of the incubation period and increased amounts of Hi.

The nucleosides adenosine, guanosine and inosine, like glucose, diminish the rate of autohaemolysis when added to blood. Remarkably, adenosine triphosphate (ATP) strikingly retards haemolysis in PK deficiency, although glucose itself is ineffective.¹⁸ ATP does not pass the red cell membrane.

The autohaemolysis test lacks specificity. This has drawn much criticism on the test, including the suggestion that it has no place in the screening of blood for inherited defects.¹⁹ The best way to detect metabolic defects in red cells is undoubtedly to measure glucose consumption, lactate production and the contribution to metabolism of the pentose phosphate pathway. These measurements are, unfortunately, difficult and are likely to be undertaken only by specialized laboratories. The autohaemolysis test does provide some information about the metabolic competence of the red cells and helps to distinguish membrane defects from enzyme defects.

In summary, we feel that the autohaemolysis test is still useful in the investigation of patients who have or who may have chronic haemolytic anaemia for the following reasons:

1. If the result is entirely normal, an intrinsic red cell abnormality is unlikely.

2. If abnormal haemolysis is fully corrected by glucose, a metabolic abnormality is unlikely and a membrane abnormality is likely.

3. If abnormal haemolysis shows little or no correction by glucose, a metabolic abnormality is likely, provided that obvious features of spherocytosis are not present on the blood film.

Thus, in our experience, a combination of red cell morphology with the results of the autohaemolysis tests makes it possible to differentiate membrane abnormalities from enzyme deficiencies in the vast majority of cases.