Pitfalls in the diagnosis of autoimmune haemolytic anaemia

The diagnosis of autoimmune haemolytic anaemia (AIHA) is usually simple and based on the presence of haemolytic anaemia and serological evidence of anti-erythrocyte autoantibodies, detected by the direct antiglobulin test (DAT)¹. There are several DAT methods with different sensitivities and specificities, levels of automation and diffusion among laboratories: the DAT-tube method is the traditional agglutination technique, usually performed with broad-spectrum Coombs’ reagents. Clinically, it is important to perform the test with monospecific antisera, to distinguish “warm”, “cold”, and “mixed” forms. The first (~70% of all cases) are mainly due to immunoglobulin (Ig) G which generally react at 37 °C, are usually directed against epitopes of the Rhesus (Rh) system, and primarily determine extravascular haemolysis. Cold forms (~20% of all cases) are due to IgM, which are able to fix complement more efficiently than other isotypes, have an optimal temperature of reaction at 4 °C, are directed against the I/i system, and prevalently cause intravascular haemolysis; cold IgM autoantibodies can easily be detected by the spontaneous agglutination of red blood cells (RBC) at 20 °C. Of note, the amount of erythrocyte destruction by intravascular haemolysis has been calculated as 200 mL of RBC in 1 hour, whereas the destruction by extravascular haemolysis is 10-fold less. According to DAT results and to the thermal characteristics of the autoantibody, AIHA are usually classified into warm forms (DAT+ for IgG only or IgG plus C3d), cold agglutinin disease (DAT+ for C3d only, with cold agglutinins of I specificity), and mixed forms (DAT+ for IgG and C3d, with coexistence of warm autoantibodies and high titre cold agglutinins). It is worth mentioning the Donath-Landsteiner autoantibody, a bithermic haemolysin able to fix complement at cold temperatures and to cause RBC lysis at 37 °C, directed against the erythrocyte P antigen, and responsible for paroxysmal cold haemoglobinuria, a rare disease mainly observed as an acute form in children (<1% of all AIHA, ~30% of paediatric cases)^1,2.

It is important to remember that DAT-tube with polyspecific or anti-IgG and anti-C antisera may yield false-negative results because of the presence of IgA, low-affinity autoantibodies, or numbers of RBC-bound IgG molecules below the threshold of the test (400 molecules per RBC). For the former two conditions, the use of monospecific antisera against IgA and low ionic strength solutions (LISS), or cold washing can overcome the DAT negativity. Small amounts of RBC-bound IgG can be detected by employing more sensitive but less specific techniques such as microcolumn and solid-phase antiglobulin tests, which are suitable for automation and are nowadays the most commonly used methods. As regards more sophisticated techniques that are not routine in the majority of laboratories, it is worth mentioning the complement-fixation antibody consumption test, enzyme-linked and radiolabeled tests, and flow-cytometry, which has the highest sensitivity, being able to detect up to 30–40 molecules of anti-RBC autoantibodies. Moreover, the dual DAT³ is useful for the detection of RBC-bound IgM warm antibodies, which fail to be revealed by standard methods, and can cause severe or fatal AIHA⁴. Finally, mitogen-stimulated-DAT (MS-DAT)⁵ is a functional and quantitative method for the detection of anti-RBC antibodies in whole blood cultures stimulated with mitogen, which amplifies the production of autoantibodies and allows the detection of small amounts of RBC-bound IgG. The test has been proven to be useful in the diagnosis of DAT-negative AIHA⁶, and found positive in a fraction of patients with B-chronic lymphocytic leukaemia⁷ or myelofibrosis⁸ without an overt diagnosis of AIHA, suggesting that in vitro mitogen stimulation could disclose a latent anti-RBC autoimmunity.

Recently two cases of IgA-positive AIHA were described. Bajpayee et al.⁹ reported a case secondary to Hodgkin’s disease in which the DAT was negative when performed by the conventional tube technique using polyspecific anti-human globulin, column agglutination technique, and after washing the RBC in cold saline; the DAT became positive when performed using a monospecific antiglobulin gel card with IgA antiglobulin. Villa et al.¹⁰ described a case of AIHA in a pregnant woman, in whom the traditional DAT-tube was negative when performed with polyspecific, monospecific anti-IgG and anti-C3 antisera from three manufacturers, and anti-IgA and IgM antisera from one manufacturer. The DAT became positive when performed with a gel column test with anti-IgA antiglobulin, confirming the greater sensitivity of this method, compared with traditional DAT-tube. At variance, more sensitive techniques, such as the solid-phase and the mitogen-stimulated DAT, gave negative results. For the latter test we had described a case of IgA-positive AIHA in which mitogen stimulation in vitro also induced IgG production⁶; however, this is not a general rule, depending on the prevailing B lymphocyte subset stimulated in vitro. In a comparative study of various DAT methods we found that DAT-tube was the most specific but least sensitive test (0.87 and 0.43, respectively); other traditional DAT methods (microcolumn/solid phase) showed reduced specificity but increased sensitivity (0.70 and 0.65, respectively), and MS-DAT was the least specific but the most sensitive test (0.59 and 0.88, respectively), underlying that the counterpart of a greater sensitivity is a reduced specificity⁶. In fact, it has been reported that 0.01–0.1% of healthy blood donors and 0.3–8% of hospital patients have a positive DAT without clinical evidence of AIHA. Moreover it should be remembered that the DAT may be positive after administration of various therapeutics (intravenous immunoglobulins, Rh immune globulins, antilymphocyte globulin and antitymocyte globulin), and in diseases with elevated serum globulins or paraproteins. Finally, the DAT is positive in conditions such as delayed haemolytic transfusion reactions caused by alloantibodies, and in haemolytic disease of the newborn; in these cases the correct interpretation of a positive DAT requires the search for irregular antibodies (indirect antiglobulin test) in the serum and/or eluate prepared from the patient’s red blood cells, and their identification if present¹. Altogether these findings and the case by Villa et al.⁹ reinforce the assertion by Petz and Garratty that “no one test was found to be optimal; a battery of tests seems to be the most efficient approach to the diagnosis of DAT-negative AIHA”¹. However, despite the numerous tests available for the detection of antibodies against RBC, and the development of additional more sensitive techniques, about 10% of AIHA remain DAT negative, and the diagnosis is made after extensive laboratory investigations to exclude other causes of haemolysis, and on the basis of the clinical response to therapy¹. These forms, reported with increasing frequency in both adults and children, represent a critical diagnostic problem and may cause harmful delays in therapy¹¹.

In a recent paper¹² we described the clinical heterogeneity and the predictors of outcome in 308 cases of primary AIHA, showing that mixed, atypical (mainly DAT-negative) and warm IgG+C AIHA (~30% of cases) more frequently have a severe onset with haemoglobin levels <6 g/dL and require multiple lines of therapy. Moreover, infections, particularly after splenectomy, acute renal failure, Evans’ syndrome, and multi-treatment were predictors of a fatal outcome. In this large multicentre study, DAT was firstly performed by standard tube or microcolumn with polyspecific (anti-IgG+C) and monospecific anti-IgG, and anti-C3 antisera. In negative cases the DAT was carried out also with monospecific anti IgA and anti-IgM antisera, after washing RBC with 65% LISS, and in solid-phase. DAT-negative cases were further investigated by MS-DAT. As shown in Table I, we identified 16 atypical cases: one DAT positive for IgA only, six positive only by MS-DAT, and nine DAT-negative after all the investigations performed. The median haemoglobin value at onset was 6.2 g/dL (range, 3–9 g/dL), with half of the patients presenting with a severe clinical picture. Lactate dehydrogenase was increased in almost all cases (median 816 U/L; range, 323–8,681 U/L) indicating considerable intravascular haemolysis, reticulocytopenia/inadequate reticulocytosis (expression of poor bone marrow compensation) was present in 6/14 (43% of cases), and thrombocytopenia (Evans’ syndrome) in two patients. As regards therapy, all cases were treated with steroids (mostly parenteral), eight received transfusions, and seven patients required second-line treatment (5 immunosuppressants/rituximab and 2 splenectomy). It is worth commenting that first-line treatment with steroids is usually prescribed without particular concern in DAT-negative AIHA, and may also be useful to clarify the diagnosis; by contrast, the indication for more aggressive and potentially harmful second-line therapy may be troublesome when the diagnosis is uncertain. In these cases, any further non-routine investigations (MS-DAT, ELISA, flow-cytometry) may be of great value in helping clinical decisions and should be strongly pursued.

Table I.

Clinical and haematological data of 16 DAT (tube, microcolumn and solid phase)-negative AIHA.

Patient N.	Gender	Age (years)	Hb (g/dL)	LDH (U/L)	Reticulocytes ×103/mmc	Serological findings
1	F	68	5.7	793	550	DAT IgA+ only
2	M	42	5.9	1,867	130	MS-DAT+ only
3	F	45	6	323	100	MS-DAT+ only
4	M	39	9	700	260	MS-DAT+ only
5	F	42	6.8	560	187	MS-DAT+ only
6	M	46	7.5	625	258	MS-DAT+ only
7	F	31	3	2,000	300	MS-DAT+ only
8	F	49	5.6	888	99	All negative
9	F	46	4.1	360	120	All negative
10	M	57	5.8	8,681	29	All negative
11	F	33	7.3	561	195	All negative
12	F	27	8	400	275	All negative
13	F	73	9	1,032	119	All negative
14	F	25	6.4	1,631	780	All negative
15	F	44	7.6	1,698	279	All negative
16	M	45	5.4	839	202	All negative

Ranges of normal values: haemoglobin (Hb) 12–16 g/dL for females and 13.5–17.5 g/dL for males; lactate dehydrogenase (LDH): 135–214 U/L; reticulocytes: 16–84×10³/mmc.

In conclusion, the percentage of “true” DAT-negative cases in our series (~3%, 9/308) is lower than that reported by Petz & Garratty¹, indicating that an extensive investigation can reduce the fraction of DAT-negative cases, avoiding delayed diagnosis and inappropriate therapies. It should be emphasised that a tight collaboration and continuous dialogue between clinicians and immunohematology laboratory experts is warranted to offer the best diagnosis and therapy for atypical cases of AIHA.