Review: Acute Donath-Landsteiner hemolytic anemia

Full text

The term paroxysmal cold hemoglobinuria (PCH)
was first used in the early 1900s to describe immune
hemolysis caused by the Donath-Landsteiner (DL)
antibody, an IgG antibody with anti-P specificity that
mediates biphasic hemolysis. The disease was aptly
named for a recurrent complication in late-stage or
congenital syphilis, observed as sudden attacks
(paroxysms) of constitutional symptoms and
hemoglobinuria precipitated by exposure to cold
temperatures.1(pp12–3) Today, the diagnosis of chronic
PCH in adults is extremely rare, its decline prompted by
the advent of effective treatment for syphilis (Table 1).
In contrast, the presence of DL antibodies should be
immediately suspected in children with the abrupt
onset of intravascular hemolysis that follows a recent
viral or bacterial illness.2–4 In children, the symptoms
are not induced by cold exposure, resolve completely
within a few weeks, and do not recur. Because the
clinical presentation of acute disease in children differs
from that of chronic PCH in adults, the condition is
alternately referred to as Donath-Landsteiner hemolytic
anemia (DL-HA).3Overall, immune hemolysis is an
uncommon diagnosis in pediatrics, but DL antibodies
are disproportionately implicated in 30 to 40 percent
of cases of autoimmune hemolytic anemia in young
children.2–4 This review summarizes the clinical and
serologic aspects of the acute transient form of
hemolysis caused by DL antibodies in children.
Pathogenesis
Julius Donath and Karl Landsteiner in 1904 were
the first to attribute the temperature-dependent
hemolysis in PCH to a cold-reacting autohemolysin and
warm-reacting lytic factor.1(pp12–3) Their test for biphasic
hemolysis after a cold-to-warm transition was the first
immunohematologic test ever described and still
serves as the definitive test for the Donath-Landsteiner
autoantibody.1(pp12–3) The DL autoantibody is an IgG
antibody that sensitizes RBCs at cold temperatures
(< 20°C) by fixing the early components of comple-
ment, then dissociates from the RBCs at warmer
temperatures. Complement activation is maximal at
37°C and proceeds to completion to produce
hemolysis only after the warm incubation. Although
initial incubation at low temperatures (0°C–20°C) is
required for hemolysis in vitro, DL antibodies are also
capable of sensitizing RBCs under physiologic
conditions, even though peripheral body temperature
rarely falls below 30°C. The mechanism by which DL
antibodies produce such severe intravascular hemoly-
sis is incompletely understood. Possible explanations
include a heterogeneous population of DL antibodies
with differential affinity in vivo and the ability of DL
antibodies to initiate repeated cycles of complement
activation by binding, dissociating, then reattaching to
unaffected RBCs as blood circulates from cooler
peripheral temperatures to warmer core temper-
atures.2,3 In addition to the intrinsic characteristics of
the DL antibody, patient-related factors, such as
concurrent infection, may also contribute to an
increased susceptibility to complement-mediated
hemolysis in DL-HA.
The inciting stimulus for autoantibody formation is
unknown, but infection or immune dysregulation may
alter pathways that otherwise would suppress
formation of antibodies against self-antigens.5
Alternatively, pathogens may alter the RBC membrane
to stimulate autoantibody formation or may possess
antigens similar to those on RBC membranes so that
56 IMMUNOHEMATOLOGY, VOLUME 21, NUMBER 2, 2005
Review: acute Donath-Landsteiner
hemolytic anemia
A.F. EDER
Table 1. Classification of DL antibodies in a single institution2
Number
Clinical classification Age (percent)
Acute Children (< 13 years old) 30 (58%)
(Acute transient
nonsyphilitic PCH) Adult (> 27 years old) 14 (27%)
Chronic, nonsyphilitic PCH Adult 3 (6%)
Chronic, syphilitic PCH 77 years old 1 (2%)
Unknown; incidental 67–82 years old 4 (7%)
Total = 52

IMMUNOHEMATOLOGY, VOLUME 21, NUMBER 2, 2005 57
Donath-Landsteiner anemia
the resultant antibodies cross-react with RBCs (e.g.,
molecular mimicry).6DL-HA often occurs 1 to 3 weeks
after a viral or bacterial infection and a wide variety of
pathogens have been implicated in case reports (Table
2).1(pp73–9)–4 More commonly,a specific etiologic agent is
not identified, but an upper respiratory tract infection
precedes the hemolysis in a vast majority of cases.2,3
DL antibodies have specificity for the P antigen, a
glycosphingolipid globoside that is also commonly
expressed on microorganisms. The erythrocyte P
antigen is the cellular receptor for parvovirus B19, and
individuals who lack the P antigen (p or Pkpheno-
types) are resistant to infection with the virus.7,8 Acute
infection with parvovirus B19 has been described in a
child with DL-HA, but it does not appear to be a
primary or underrecognized cause of DL-HA.9,10 In a
small case series of patients with DL antibodies, none
had IgM antibodies against parvovirus B19, and IgG
antibodies were not more common in the patients (7 of
13) compared to the control group (11 of 18).10 Finally,
several temporal clusters of acute DL-HA have been
reported, but a common etiologic agent was not
implicated in the cases.2,4 Apparently, a wide variety of
infections can trigger production of DL antibodies;
moreover, increased awareness of the condition likely
results in improved detection.4Rare examples of DL
antibodies have been described in patients, typically
adults, with other immunologic disorders, such as
lymphoproliferative malignancies, collagen disease,
myelodysplastic syndrome, delayed hemolytic
transfusion reaction, and other types of autoimmune
hemolytic anemia.2,11–13
Clinical Presentation and Management
The typical presentation of DL-HA is a young child
with a recent history of an upper respiratory tract
infection or other acute illness associated with
recurrent fever, and the passage of red-brown urine.
Common characteristics in childhood cases from the
two largest observational series describing DL
antibodies are summarized in Table 3. DL-HA most
often occurs in children under the age of 5 years
(range,8 months to 13 years).2,3 In some reports, there
is a slight male preponderance, with a male to female
ratio of about 2:1.2–4 Overall, DL-HA is an uncommon
condition, but it may account for 30 to 40 percent of
cases of autoimmune hemolytic anemia in young
children. 2–4 The precise incidence and prevalence of
DL-HA is unknown, and estimates based on small
numbers of cases reported to blood services or
reference laboratories, while informative, should be
interpreted cautiously. Sokol et al. estimated the annual
incidence of DL-HA as 0.4 per 100,000 for children
under the age of 5 years.2The transient nature of the
DL antibody, the level of awareness among primary
care physicians, and the availability of the DL antibody
are factors that potentially affect recognition and
correct diagnosis of DL-HA.
The sudden onset of hemoglobinuria is reported in
almost all cases of acute PCH, often accompanied by
pallor, jaundice, and fever. The intravascular hemolysis
and a rapidly progressing anemia may have a dramatic
clinical presentation as high fever, shaking chills, and
abdominal pain. Headache, nausea, vomiting, anorexia,
and diarrhea may also occur. Physical findings included
an enlarged liver and spleen in 25 percent of cases.4
Typically, symptoms in children with acute DL-HA are
not precipitated by cold exposure, in contrast to the
presentation of chronic, syphilitic PCH in adults.
The degree of anemia is variable but may be severe
in children with DL-HA. About one third of patients
have hemoglobin concentrations of 5 g/dL or less
(range: 2.5–12.5 g/dL) at presentation, which may
decrease rapidly in the first 12 to 24 hours.2–4
Differences in the severity of anemia among patients
may reflect the characteristics of the DL antibody, such
as serum titer or thermal range, or the interval of time
between the onset of hemolysis and diagnosis. In
addition, reticulocytopenia is often observed early in
the course of the disease and the delayed
hematopoietic response aggravates the initial anemia.
Reticulocytopenia may result from viral suppression of
the bone marrow or preferential destruction of
reticulocytes by the DL autoantibody. However, the
period of reticulocytopenia usually is brief and
reticulocytosis commensurate with the degree of
hemolysis promptly ensues. The peripheral blood
Table 2. Pathogens implicated in case reports of acute DL-HA
Upper respiratory tract infection
Gastroenteritis, enteritis
Measles
Mumps
Chicken pox
Cytomegalovirus (CMV)
Epstein-Barr virus
Influenza virus
Adenovirus
Parvovirus B19
Coxsackie virus A9
Haemophilus influenza
Mycoplasma pneumoniae
Klebsiella pneumoniae
Measles vaccine

58 IMMUNOHEMATOLOGY, VOLUME 21, NUMBER 2, 2005
A.F. EDER
smear demonstrates RBC agglutination, polychromasia,
nucleated RBCs, anisopoikilocytosis, occasional
spherocytes, and erythrophagocytosis (Fig. 1). Erythro-
phagocytosis by neutrophils, rather than monocytes, is
a relatively frequent phenomenon in PCH,but is rarely
observed in other types of autoimmune
hemolytic anemia.14 The significance of this
phenomenon to the RBC destruction that
occurs is unclear, but extravascular hemolysis
may contribute to the anemia and is further
evidenced by occasional spherocytes on the
peripheral blood smear. In addition, the
monocyte monolayer assay (MMA) performed
with mononuclear cells from patients with
PCH demonstrated significantly more
phagocytic activity than with those from
normal individuals, suggesting enhanced
reticuloendothelial cell function in the
disease.4The practical significance of
erythrophagocytosis, especially by neutro-
phils, in a young child is that it should trigger
further investigation for DL-HA.4Additional
serum chemistry findings in patients with DL-
HA include increased LDH and unconjugated
(indirect) bilirubin, decreased or absent
haptoglobin, increased blood urea nitrogen
(BUN), and increased serum creatinine.
Acute DL-HA usually resolves spontane-
ously and completely within several weeks
and does not recur. There is a single case
report of acute PCH occurring on two separate
occasions in a child, each time after an upper
respiratory tract infection, although the specificity of
the antibody was not anti-P.15 Treatment is supportive;
glucocorticoids do not shorten the course of the
disease. If glucocorticoids were started empirically for
treatment of AIHA, they can be discontinued once the
diagnosis of DL-HA is confirmed. Transfusion may be
needed if anemia is severe.
DL antibodies do not interfere in routine
pretransfusion and compatibility tests, because the
causative autoantibody rarely causes RBC agglutination
above 20°C. Patients with DL-HA typically demonstrate
a DAT that is positive for complement, but negative for
IgG, with a corresponding eluate that is negative. The
antibody screen is usually negative, because DL
antibodies do not react with RBCs under routine
reaction conditions. DL antibodies may cause direct
agglutination at 0°C or may be detected if the IAT is
performed under strict conditions in the cold.
Although the specificity of the DL antibody is
almost always anti-P, P– RBCs (i.e. p, P1
k, or P2
k)are
extremely rare and are not available in routine practice
for patients who require urgent transfusion.
Fortunately, most patients with DL-HA who require
Table 3. Common findings in DL-HA in children
Gottsche et al., 1990 Sokol et al., 1999
Number of cases (duration of study) 22 cases (4 years) 30 cases (37 years)
Patient age mean (range) 2.6 years (8 months–5 years) 3.5 (1–13 years)
Clinical presentation
Percent (number) with precedent URI 77% (17/22) URI 87% (26/30)
(< 3 weeks) infection Other 23% (5/22) Other 7% (2/30)
None 7% (2/30)
Percent (number) with severe anemia 27% (6/22) 40% (12/30)
(Hb < 5 g/dL)
Hemoglobin (g/dL)(range) 6.1 (4.4–8.8) 6.0 (3.4–12.9)
Percent (number) with hemoglobinuria 77% (17/22 ) 100% (30/30)
Serologic characteristics
DAT
Complement only 100% (22/22,at 37°C) 90% (27/30)
73% (16/22, at 20°C)
Complement + IgG 27% (6/22,at 20°C) 7% (2/30)
Not tested – 3% (1/30)
Positive DL test
Untreated RBC 59% (13/22) 93% (28/30)
Enzyme-RBC, only 41% (9/22) 3% (1/30)
Two-stage,only – 3% (1/30)
Anti-P, if evaluated NR 13/13 tested
URI = upper respiratory tract infection
NR = not reported
Fig. 1. Erythrophagocytosis in DL-HA.Peripheral blood from a 2-year-old
girl with DL-HA and sudden onset of severe anemia (Hb 4.2
g/dL), showing erythrophagocytosis and RBC agglutination
(100×;Wright stain).

IMMUNOHEMATOLOGY, VOLUME 21, NUMBER 2, 2005 59
Donath-Landsteiner anemia
transfusion achieve a favorable clinical response and an
adequate posttransfusion increment with P+ RBCs
despite the presumed, or demonstrated, incompat-
ibility. P– RBCs from rare donor registries have been
used in case reports of PCH when hemolysis was
severe and prolonged, although it is not clear whether
transfusion of P– RBCs was beneficial or spontaneous
recovery was coincident with transfusion.16 Although
cold-induced hemoglobinuria is rarely a feature of DL-
HA, the patient should be kept warm during the
transfusion and the use of a blood warmer is prudent
despite the paucity of data that support this practice.
Washing RBCs to remove residual complement may
have theoretical benefit, but likely does not improve
transfusion safety and often is not performed.1(p396) In
rare cases of life-threatening anemia, plasmapheresis
has been used to acutely remove IgG autoantibodies
and alleviate symptoms.17
Donath-Landsteiner Antibodies
Donath-Landsteiner antibodies are typically IgG
with anti-P specificity that demonstrate a low titer
(< 32), a low thermal amplitude (< 20°C), and biphasic
hemolysis. The vast majority of DL antibodies
have these characteristics; however, the following
exceptions have been described in case reports:
• Specificity: Biphasic IgG antibodies may
demonstrate specificity for antigens other than P,
such as anti-I, anti-p (anti-Gd), anti-i, and anti-“Pr-
like.”1(p254)
•Thermal activity: Biphasic IgG antibodies with
anti-P specificity may demonstrate thermal
activity above 20°C, agglutinate RBCs, or react by
IAT at 37°C.1(pp223–7),18–21
•Immunoglobulin subclass: IgM antibodies may
give a positive DL test,but these usually represent
falsely positive results caused by monophasic
hemolysis. Regardless, patients demonstrating
IgM antibodies with biphasic hemolytic
properties have been diagnosed as having PCH.
The antibodies had anti-I, anti ITP, or anti-P
specificity.1(pp195-6)
Although there is no established consensus,
diagnosis of DL-HA should require, at a minimum, a
biphasic IgG antibody even if the specificity is not anti-
P and evidence of intravascular hemolysis.1(pp195–6) The
diagnosis of DL-HA should be questioned when the
clinical history is atypical or the antibody is IgM or has
other characteristics unusual for DL antibodies.
The IgG class of DL antibodies can be confirmed by
performing the DAT under cold conditions, including
cold washes to avoid eluting antibody. In addition,IgG
may be demonstrated if IAT is performed in the cold
with monospecific antihuman globulin reagents,
although this test is susceptible to interference from
normal cold agglutinins. In one report, a DL antibody
was identified as IgG3, although it is not known
whether this is a consistent finding in cases of DL-HA.22
Finally, the MMA may be strongly positive in patients
with PCH, suggesting the presence of IgG on the
surface of their RBCs, and may be more sensitive than
the IAT.4,23
Donath-Landsteiner Assay for Biphasic
Hemolysis
The essential test for DL-HA is the Donath-
Landsteiner assay for biphasic hemolysis,which can be
performed either by the direct method using a sample
of whole blood or by the indirect method using
separated serum.1(pp223–7),2 The patient’s blood
specimens must be maintained at 37°C after collection.
Both assays first require an incubation of sample tubes
in a melting ice bath (approximately 0°C) followed by
an incubation at 37°C. The endpoint is the presence or
absence of visible hemolysis in the plasma compared to
control tubes kept at a constant temperature (0°C or
37°C). The direct test is an easy screening test but
suffers from several limitations, as it requires more
whole blood and is less sensitive than the indirect test.4
Autologous RBC lysis may not be observed because of
the protective effect of C3dg deposited on the surface
of circulating RBCs during the hemolytic episode.
Moreover, recent or ongoing hemolysis may deplete
serum complement so that RBC lysis is not detected in
the direct assay (Table 4).
The indirect test is performed with the patient’s
serum which has been maintained and promptly
Table 4. Limitations of DL tests
False-negative results
Direct DL test (lysis of autologous RBCs)
Low antibody titer
Low serum complement
Resistance to lysis due to C3dg
Indirect DL test (lysis of reagent RBCs)
Low antibody titer
Inhibition of DL antibody by globoside in normal serum
Autoadsorption of antibody during serum separation
False-positive results
Direct and indirect DL test
Monophasic lysis by an IgM autoantibody

separated at 37°C. A common method used for the DL
indirect assay requires three sets of three tubes (Fig.
2).24 The first set contains the patient’s serum (row 1),
the second set contains the patient’s serum and fresh
normal serum as an added source of complement (row
2), and the third set contains fresh normal serum as a
control (row 3). Reagent ABO-compatible RBCs are
added to all tubes. One tube from each set is incubated
in a melting ice bath (approximately 0°C), then
transferred to 37°C. The control tubes in each set are
kept at a constant temperature, either 37°C or 0°C.
Visible hemolysis in the patient’s samples, with or
without the additional complement,and the absence of
hemolysis in all control tubes is a positive test result.
Although maximal hemolysis is observed when the
initial incubation is 0°C, DL antibodies in acute PCH
may have thermal activity up to 24°C, but rarely cause
lysis with higher temperatures in the initial incubation.
If RBCs with the p or Pkphenotype are available,
they can be used as a negative control and should
remain intact in the assay, confirming the expected
anti-P specificity of the DL antibody. If an antiglobulin
test is performed on the unlysed cells,antibody binding
to P+ but not to p or PkRBCs will be demonstrated.
Because these RBCs are often not available, this
additional testing may not be possible.
The most common cause of negative DL tests in
suspected cases of DL-HA, with a classic clinical
presentation and otherwise consistent laboratory
findings, is the failure to detect the transient
autoantibody that disappears quickly from the plasma
during recovery from the acute illness. DL antibody
titers quickly wane soon after recovery from the initial
hemolytic episode. False-negative results may occur in
both the direct and indirect test owing to low antibody
titers. Because the indirect test evaluates the patient’s
serum only, false negatives may result from
autoabsorption of antibody during serum separation or
inhibition of antibody by neutralizing carbohydrate
antigens (e.g., globoside) present in the added fresh
serum (Table 4).
If the diagnosis of DL-HA is strongly suspected, but
the indirect DL test is negative, the biphasic hemolysis
test can be modified to increase its sensitivity by
treating the reagent RBCs with enzymes,by performing
a two-stage assay, or by testing for the DL antibody by
the IAT. In addition, interpretation of the DL test may be
difficult if the initial sample drawn for analysis is
hemolyzed. In these cases, the extent of additional
hemolysis during the assay may be gauged by the size of
the red cell button after centrifugation. Alternatively,
the two-stage test can be performed. In rare patients
with classic DL-HA, the DL test was only positive with
one of the following modified tests (Table 3)2,3:
•Enzyme-treated RBCs: The reagent RBCs may be
treated with enzymes, such as 1% papain, which
exposes more P antigen on the RBC membrane.
Enzyme treatment, however, makes the RBCs
more susceptible to lysis by cold agglutinins, and
careful interpretation of appropriate controls is
necessary.
•Two-stage test: For the two-stage test, the
patient’s serum is replaced with normal ABO-
compatible serum after the first incubation at
0°C. Because antibody binding to the RBC occurs
Ice Ice 37°C
then
37°C
Reaction conditions
60 IMMUNOHEMATOLOGY, VOLUME 21, NUMBER 2, 2005
A.F. EDER
Fig. 2. Indirect Donath-Landsteiner test: sample conditions.
1ABC
Set 1: Patient’s serum
2
Set 2: Patient’s serum + Normal serum
3
Set 3: Patient’s serum + Normal serum
ABC
ABC
ABC

IMMUNOHEMATOLOGY, VOLUME 21, NUMBER 2, 2005 61
only in the first phase, the patient’s serum can be
removed after this incubation occurs. Serum
replacement not only removes the patient’s
serum that may have been hemolyzed at baseline,
but also provides additional complement for the
second phase of the assay at 37°C without
diluting the sample.
•IAT: DL antibodies may be detectable by the IAT
with antihuman globulin after the cold
incubation with monospecific reagents.1(p224)
However, interpretation of the test is problematic
because the DL antibody or coincidental IgM
antibody may cause direct agglutination at 0°C,
and normal incomplete cold antibody may give
false-positive results. A negative control with p or
Pkphenotype RBCs should be used because of
the propensity for false-positive results.
Conclusion
The diagnosis of DL-HA is readily made in most
cases, but requires an index of clinical suspicion to
initiate investigation, because the DL test is not
performed routinely by most transfusion services.
Despite the excellent reviews that have appeared in
recent journals, the diagnosis is often still delayed,
which confounds demonstration of the pathog-
nomonic DL antibodies. A young child with a
precedent viral illness and the sudden onset of
intravascular hemolysis strongly suggests a diagnosis of
DL-HA, even in light of a negative DL test. The most
common cause of negative DL tests in acute cases is
the failure to detect the transient autoantibody that
disappears quickly from the plasma soon after the
initial hemolytic episode. Drug-induced hemolysis
(e.g., ceftriaxone), or possibly transfusion-associated
hemolysis, may be associated with acute intravascular
red cell destruction like acute DL-HA, but can be
distinguished by the clinical history and immuno-
hematologic findings.2,25 On occasion, a cold-reactive
IgM autoantibody may demonstrate apparent biphasic
behavior; conversely, a DL antibody may appear to be
monophasic because its thermal amplitude is close to
physiologic temperatures. DL-HA is differentiated from
cold agglutinin syndrome on clinical grounds and
laboratory findings.1(p196) Although a rare disease, and
rarely a diagnostic dilemma, DL-HA nonetheless
requires prompt recognition for laboratory diagnosis
and appropriate clinical management.
References:
1. Petz LD, Garratty G. Acquired immune hemolytic
anemias. 2nd ed. New York: Churchill Livingstone,
2004.
2. Sokol RJ, Booker DJ, Stamps R. Erythropoiesis:
paroxysmal cold haemoglobinuria: a clinico-
pathological study of patients with a positive
Donath-Landsteiner test. Hematology 1999;4:
137-64.
3. Gottsche B, Salama A, Mueller-Eckhardt C. Donath-
Landsteiner autoimmune hemolytic anemia in
children. A study of 22 cases. Vox Sang 1990;58:
281-6.
4. Heddle NM. Acute paroxysmal cold hemoglobin-
uria.Transfus Med Rev 1989;3:219-29.
5. Mackay IR. Science, medicine and the future:
tolerance and autoimmunity. BMJ 2000;321:93-6.
6. Oldstone MB. Molecular mimicry and immune-
mediated diseases. FASEB J 1998;12:1255-65.
7. Brown KE,Anderson SM Young NS. Erythrocyte P
antigen: cellular receptor for B19 parvovirus.
Science 1993;262:114-7.
8. Brown KE, Hibbs JR, Gallinella G, et al. Resistance
to parvovirus B19 infection due to lack of virus
receptor (erythrocyte P antigen) N Engl J Med
1994;330:1192-6.
9. Chambers LA, Rauck AM. Acute transient hemo-
lytic anemia with a positive Donath-Landsteiner
test following parvovirus B19 infection. J Pediatr
Hematol Oncol 1996;18:178-81.
10. Sharpe JS, Booker DJ, Stamps R, Sokol RJ.
Parvovirus B19 infection and paroxysmal cold
hemoglobinuria. Transfus Med 1996;6
(Suppl 2):24.
11. Sivakumaran M, Murphy PT, Booker DJ, et.al.
Paroxysmal cold haemoglobinuria caused by non-
Hodgkin’s lymphoma. Br J Haematol 1999;105:
278-9.
12. Breccia M, D’Elia GM, Girelli G, et al. Paroxysmal
cold hemoglobinuria as a tardive complication of
idiopathic myelofibrosis. Eur J Haematol 2004;73:
304-6.
13. Wodsinski MA, Collin RC, Booker DJ, et al. Delayed
hemolytic transfusion reaction and paroxysmal
cold hemoglobinuria: an unusual association.
Immunohematol 1997;13:54-7.
14. Garratty G.Erythrophagocytosis on the peripheral
blood smear and paroxysmal cold hemoglob-
inuria.Transfusion 2001;41:1073-4.
Donath-Landsteiner anemia

62 IMMUNOHEMATOLOGY, VOLUME 21, NUMBER 2, 2005
15. Taylor CJ, Neilson JR, Chandra D, Ibrahim Z.
Recurrent paroxysmal cold haemoglobinuria in a
3-year-old child: a case report.Transfus Med 2003;
13:319-21.
16. Rausen AR, Levine R, Hsu TCS, Rosenfield RE.
Compatible transfusion therapy for patients with
paroxysmal cold hemoglobinuria. Pediatrics 1975;
55:275-8.
17. Roy-Burman A, Glader BE. Resolution of severe
Donath-Landsteiner autoimmune hemolytic
anemia temporally associated with institution of
plasmapheresis. Crit Care Med 2002;30:931-4.
18. Ries CA, Garratty G, Petz LD, Fudenberg HH.
Paroxysmal cold hemoglobinuria: report of a case
with an exceptionally high thermal range Donath-
Landsteiner antibody. Blood 1971;38:491-9.
19. Lindgren S, Zimmerman S, Gibbs F, Garratty G: An
unusual Donath-Landsteiner antibody detectable
at 37 degrees C by the antiglobulin test.
Transfusion 1985;25:142-4.
20. Nordhagen R: Two cases of paroxysmal cold
hemoglobinuria with a Donath-Landsteiner
antibody reactive by the indirect antiglobulin test
using anti-IgG.Transfusion 1991;31:190-1.
21. Sabio H, Jones D, McKie VC. Biphasic hemolysin
hemolytic anemia:reappraisal of an acute immune
hemolytic anemia of infancy and childhood.Am J
Hematol 1992;39:220-2.
22. Gelfand EW, Abramson N, Segel GB, Nathan DG.
Buffy-coat observations and red-cell antibodies in
acquired hemolytic anemia. N Engl J Med 1971;
284:1250-2.
23. Garratty G, Nance S, Arndt P, et al. Positive direct
monocyte monolayer assays associated with
positive Donath-Landsteiner tests. Transfusion
1989;29(Suppl): 49S.
24. Brecher ME, ed. Technical Manual. 14th ed.
Bethesda, MD: American Association of Blood
Banks, 2002.
25. Garratty G. Review: drug-induced immune
hemolytic anemia—the last decade. Immuno-
hematol 2004;20:138-46.
Anne F. Eder MD, PhD, Executive Medical Officer,
Biomedical Services, National Headquarters,
American Red Cross, 2025 E Street NW, Washington,
DC 20006.
A.F. EDER.
REMEMBER: THE PASSWORD IS “2000” For www.redcross.org/pubs/immuno
Now, as a subscriber, you can enter the password, 2000, and access the back issues. That means
cover to cover! You will receive every article in total, every letter to the editor,every review, every ad, every
notice, and every literature review! All of the other services will continue to be available on the Web site,
including sending a letter to the editor, subscribing with a credit card on the secure order site, performing a
literature search, reviewing instructions for authors, and linking to other important sites. Log on now to see
this great service!

132 IMMUNOHEMATOLOGY, VOLUME 21, NUMBER 3, 2005
ERRATUM
Vol. 21, No.2, 2005; page 60
Review: acute Donath-Landsteiner hemolytic anemia
The author has informed the editors of
Immunohematology that there is an error on page 60,
Fig.2. Set 3 in Fig. 2 should read “Normal serum.”
COMMUNICATIONS CONT’D
Ice Ice 37°C
then
37°C
Reaction conditions
Fig. 2. Indirect Donath-Landsteiner test: sample conditions.
1ABC
Set 1: Patient’s serum
2
Set 2: Patient’s serum + Normal serum
3
Set 3: Normal serum
ABC
ABC
ABC