Conflicts of interest: Nothing to report
Characterization of direct antiglobulin test-negative autoimmune hemolytic anemia: A study of 154 cases
Article first published online: 21 NOV 2012
Copyright © 2012 Wiley Periodicals, Inc.
American Journal of Hematology
Volume 88, Issue 2, pages 93–96, February 2013
How to Cite
Kamesaki, T., Toyotsuji,, T. and Kajii, E. (2013), Characterization of direct antiglobulin test-negative autoimmune hemolytic anemia: A study of 154 cases. Am. J. Hematol., 88: 93–96. doi: 10.1002/ajh.23356
- Issue published online: 24 JAN 2013
- Article first published online: 21 NOV 2012
- Accepted manuscript online: 25 OCT 2012 07:43AM EST
- Manuscript Accepted: 17 OCT 2012
- Manuscript Revised: 11 OCT 2012
- Manuscript Received: 26 SEP 2012
- Ministry of Health
- Labor and Welfare of Japan
- JKA promotion funds from KERIN RACE
Direct antiglobulin test (DAT)-negative (DAT-)autoimmune hemolytic anemia (AIHA) is empirically thought to show the same clinical conditions as DAT-positive (DAT+)AIHA, with the exception of an adequate amount of red blood cell (RBC)-bound immunoglobulin (Ig)G. We investigated the clinical characteristics of DAT−AIHA in comparison with DAT+AIHA. Of the 582 patients referred to our laboratory with undiagnosed hemolytic anemia, AIHA was clinically diagnosed in 216 patients (DAT−AIHA, n = 154; DAT+AIHA, n = 62). The percentage of reticulocytes, mean corpuscular volume, RBC-IgG levels, white blood cell count, and total protein (TP) levels were significantly higher in patients with DAT+AIHA than patients with DAT−AIHA. The hemoglobin level was significantly lower in patients with DAT+AIHA. No significant differences between patients with DAT−AIHA and DAT+AIHA existed with respect to age, gender, idiopathic/secondary nature, complications such as Evans syndrome, effectiveness of steroid treatment, or survival rate at 1 year following diagnosis. Patients with DAT−AIHA required significantly lower doses of steroids for maintenance therapy. Based on multivariate analysis of idiopathic DAT−AIHA (n = 110), TP and Evans syndrome were associated with the effectiveness of steroids (adjusted odds ratio [aOR], 1.36/[0.1 g/dl]; 95% confidence interval [CI], 1.01–1.84) and survival at the 1-year follow-up (aOR, 0.1; 95% CI, 0.01–0.88). Our results indicate that patients with DAT−AIHA generally suffer milder anemia and hemolysis than patients with DAT+AIHA, respond equally well to steroids, and have comparable survival at 1-year. Am. J. Hematol. 88:93–96, 2013. © 2012 Wiley Periodicals, Inc.
Steroid therapy is the first choice in treatment of patients with autoimmune hemolytic anemia (AIHA) with 80% effectiveness . Without proper steroid therapy, patients with AIHA have a poor prognosis and a 31–53% mortality rate . Since the 1950s, investigators have used steroids to treat AIHA and reported very high effectiveness through several case-series studies [3-5]. There are few case–control studies involving AIHA, which show higher evidence than case-series studies.
The detection of red blood cell-bound immunoglobulin G (RBC-IgG) and complement by a direct antiglobulin test (DAT) remains the main serologic assay in the diagnosis of AIHA . Several methodologies have been investigated for the detection and evaluation of these autoantibodies. A DAT using the conventional tube technique (CTT) is the method most commonly used in blood centers and is still considered the gold standard . A positive DAT almost always exists in association with AIHA  and forms the basis for the serologic diagnosis of AIHA . However, it has also been shown that a negative DAT does not exclude the diagnosis of AIHA  and 1–10% of patients with AIHA have been reported to have a negative DAT [10-12]. These patients may carry a lower number of IgG molecules per RBC, yielding a negative tube DAT and in vivo hemolysis . In fact, Petz and Garratty  reported that autoantibodies were detected in the eluates from 11 of 27 patients with acquired hemolytic anemia and a negative DAT, and concentrated elutes reacted with RBC of common Rh phenotypes, but did not react with Rh-null RBC. The immunoradiometric assay (IRMA)  for RBC-IgG is a representative method to quantitatively detect RBC-IgG. We previously reported  the value of RBG-IgG (33 ± 13) in 100 healthy Japanese adults and the RBC-IgG required for DAT-positivity (335 ± 72). We also proposed that the RBC-IgG level could be useful for the diagnosis of DAT-negative (DAT−) AIHA and that the cut-off value should be 78.5 if RBC-IgG is measured before treatment .
Whether or not the clinical entity of DAT-AIHA differs from DAT-positive (DAT+) AIHA in terms of differences in the underlying pathogenic mechanism, such as the IgG subclass of autoantibodies, is controversial [17, 18]. Analogous to the treatment of patients with DAT+AIHA, steroid therapy is empirically thought to be effective in DAT−AIHA on the basis of case-series studies, including 27  and 9 cases .
Therefore, the aim of the study was to determine the characteristics of 154 DAT−AIHA in comparison with 62 DAT+AIHA and verify the effectiveness of steroid therapy in DAT−AIHA.
Patients and Methods
The study was conducted between 2003 and 2010 at the Laboratory of the Center for Community Medicine of Jichi Medical University in Tochigi, Japan after approval by the Institutional Ethics Panel Committee.
During the 8-year period, 605 samples from 582 patients throughout Japan were referred to our laboratory for quantification of RBC-IgG, based on the results of the DAT (mainly by the column method) in their laboratories. Of these samples, 78 (14%) were confirmed DAT-positive and 480 (86%) were DAT-negative in our laboratory, as shown by analysis of polyspecific DAT by CTT (Ortho Diagnostics, Rochester, NY) and monospecific DAT by CTT using anti-IgG and anti-C3d antibodies (Ortho Diagnostics), as per the manufacturer's instructions.
Sample preparation and immunoradiometric assay for RBC-IgG
Sample preparation and IRMA for RBC-IgG was performed according to a previous report . Briefly, heparinized whole blood (10 ml) samples were collected. The RBC layer was prepared by centrifuging the whole blood at 1,000 rpm for 20 min. The supernatant plasma and buffy coat were discarded. One-milliliter samples of packed RBCs were diluted in 10 ml of phosphate-buffered saline (PBS; pH 7.0, 0.15 M). The diluted RBCs were passed through a cotton-wool column to exclude neutrophils and monocytes, according to the method by Jeje et al. . The RBCs were washed four times with PBS, and the resulting RBCs (0.3 ml) were suspended in 0.4 ml of PBS.
The IRMA for RBC-IgG was performed according to the method by Jeje et al. , with some modifications. Samples of 125I-labeled goat anti-human IgG antisera derived (Du Pont, Wilmington, DE) were diluted in PBS containing 3% bovine serum albumin with a specific activity of 10,000 cpm/200 μl (Wakojunyaku, Osaka, Japan). A volume of 400 μl of the washed RBCs was incubated for 1 hr at 37°C with 200 μl of the diluted anti-human IgG. IgG beads were prepared using the methods described by Jeje et al. . Human IgG and beads (Sephasorb™ HP) were purchased from Sigma Chemical (St. Louis, MO) and Pharmacia Fine Chemicals (Uppsala, Sweden), respectively. Two hundred microliter samples of IgG beads (2 × 106) were added to the mixture of RBCs and 125I-labeled anti-human IgG and incubated at 37°C for 30 min. The RBCs were lysed by the addition of 80 μl of 20% Triton X-100 (Sigma Chemical). The beads were washed four times with 20% Triton X-100-containing PBS, and the radioactivity was measured using a gamma counter (Aroka, Tokyo, Japan). A standard curve was generated using human IgG standards (10–10,000 ng IgG/ml; Sigma Chemical). The percent inhibition of binding was plotted against each concentration of IgG. Using the standard curve, RBC-IgG levels were calculated after counting the number of RBC. Each attending physician was informed of the RBC-IgG level within 3–10 days of ordering.
Clinical diagnosis and clinical course questionnaire
One year after referral to our laboratory, follow-up investigations were performed; the attending physician used a questionnaire to assess the clinical diagnosis and course. The bases for clinical diagnosis of AIHA were in vivo hemolysis (low hemoglobin [Hb] concentration, high percentage of reticulocyte (Retic%), high indirect serum bilirubin (IDBIL) level, high lactate dehydrogenase (LDH) level, low haptoglobin (Hp) level, and/or high erythropoiesis level in bone marrow), and exclusion of other anemic icteric diseases without hemolysis, such as megaloblastic anemia, myelodysplastic syndrome, erythroid leukemia, congenital dyserythropoietic anemia, hepatobiliary diseases, and constitutional jaundice. AIHA was diagnosed by means of the DAT, steroid-reactivity, and exclusion of alloimmune hemolytic anemia and drug-induced hemolytic anemia with respect to the RBC-IgG value .
There is no formal consensus on the definition of response criteria in patients with AIHA . A response was defined as improvement in anemia (Hb > 10 mg/dl) and hemolysis after steroid treatment.
The distribution of categorical variables was compared among the groups using the X2 test or Fisher's exact test. The normality of the laboratory continuous variables was analyzed using the Kolmogorov–Smirnov test with Lilliefors significance correction. As most variables were not normally distributed, a Mann–Whitney U-test was used to determine the differences between patients with DAT−AIHA and DAT+AIHA. The backward multiple logistic regression method was used to select prognostic or therapeutic factors associated with effectiveness of steroids or survival at the 1-year follow-up evaluation in patients with idiopathic DAT−AIHA after forcing age and gender in the model; the statistical level of significance determined with the log-likelihood ratio test was set at P < 0.10. Statistical significance was defined as a two-sided P value < 0.05. All statistical computations were made using JMP statistical software 10.0.1 for Macintosh (SAS Institute, Inc., Cary, NC).
Four hundred fifteen surveys were returned for analysis, for a response rate of 71%. The mean age of the participants was 50.9 ± 25.3 years (range, 0.0–91 years [a lower age range of 0.0–0.1 contained four infants; suspected hemolysis with congenital TTP, with maternal SLE/APS, with maternal AIHA, and with CMV infection, there were no samples sent to evaluate maternal Rh Ig administration.]) and 55% of the participants were female. There were no significant differences in age and gender between the responders and non-responders (50.6 ± 25.6 or 50.7 ± 25.2 years of age and 54% or 55% female, respectively). Of the responders, 227 were diagnosed with AIHA. Sixty-four of the responders were DAT-positive; 62 had warm-type AIHA (DAT+AIHA) and 2 had cold-type AIHA. One hundred sixty-three responders were DAT-negative; 154 had warm-type AIHA (DAT−AIHA) and 9 had cold-type AIHA.
The clinical characteristics and laboratory measurements of patients with DAT−AIHA in comparison with DAT+AIHA are shown in Table 1. No significant differences between patients with DAT−AIHA and DAT+AIHA existed with respect to age, gender, idiopathic/secondary nature, or complications such as Evans syndrome. The percentage of reticulocytes, mean corpuscular volume, RBC-IgG level, white blood cell count, total protein (TP) level, and Hp were significantly higher in patients with DAT+AIHA than DAT−AIHA. Hemoglobin level was significantly lower in DAT+AIHA. No significant differences existed between patients with DAT−AIHA and DAT+AIHA with respect to the effectiveness of steroid treatment and survival rate 1 year following diagnosis (Table 2). Patients with DAT−AIHA required significantly lower doses of steroids for maintenance therapy. Patients with idiopathic warm AIHA had almost the similar pattern between DAT−AIHA (n = 110) and DAT+AIHA (n = 41; Table 3). Patients with secondary warm AIHA had also the same tendency, but a relatively poor response to treatments (Table 4). Other treatments had also been surveyed and the data of efficacy and 1-year survival following diagnosis are shown in Table 5. The sample sizes of these treatments were not large enough for statistical analyses. Treatments by intravenous gamma globulin and transfusion alone were mainly used for children (mean age, 5.2 ± 7.8 years) and young patients (mean age, 17 ± 11.8 years). In contrast, treatments with immunosuppressants were adapted for elderly patients (mean age, 71 ± 18.4 years), and treatments by splenectomy and observation were mainly used for adult patients (mean age, 57.9 ± 21.2 years) and (mean age, 56.5 ± 17.3 years), respectively. Rituximab had not been used for patients with idiopathic DAT−AIHA prior to 2010.
|Variable||DAT-negative AIHA||DAT-positive AIHA||P-value|
|Mean ± SD||Mean ± SD|
|Age||50.0 ± 25.8||50.5 ± 27.2||0.89|
|Evans syndrome (%)||38.3||38.7||0.58|
|WBC (/μL)||6401 ± 363||8741 ± 565||0.001|
|Hb (g/dL)||8.0 ± 2.4||7.3 ± 2.6||0.02|
|%Retic (%)||8.8 ± 11.3||11.2 ± 9.9||0.04|
|MCV (fL)||100.2 ± 15.9||106.3 ± 13.7||0.01|
|Platelets (103/μL)||175 ± 139||203 ± 141||0.10|
|TP (g/dL)||6.5 ± 0.9||6.9 ± 1.0||0.004|
|IgG (mg/dL)||1361 ± 174||1648 ± 241||0.34|
|LDH (U/L)||577 ± 560||717 ± 850||0.18|
|IDBIL (mg/dL)||1.8 ± 1.8||1.8 ± 1.6||0.75|
|Hp (mg/dL)||11.0 ± 2.8||17.3 ± 4.5||0.048|
|RBC-IgG||179 ± 288||1397 ± 1934||<0.0001|
|DAT-negative AIHA||DAT-positive AIHA||P-value|
|Efficacy of treatments||89%||88%||0.80|
|Treatment with steroids alone||66%||73%||0.04|
|Remission within 4 weeks||68%||57%||0.54|
|1-year survival rate following diagnosis||83%||88%||0.50|
|Treatment with steroids alone||n = 102||n = 45|
|Efficacy of steroids||91%||92%||0.99|
|Remission within 4 weeks||70%||62%||0.74|
|Maintenance steroids (< 15 mg/day)||84%||50%||0.03|
|1-year survival rate following diagnosis||84%||92%||0.27|
|DAT-negative AIHA||DAT-positive AIHA||P-value|
|Efficacy of treatments||90%||94%||0.73|
|Treatment with steroids alone||65%||78%||0.25|
|Remission within 4 weeks||72%||71%||0.54|
|1-year survival rate following diagnosis||84%||87%||0.98|
|Treatment with steroids alone||n = 71||n = 32|
|Efficacy of steroids||91%||93%||0.99|
|Remission within 4 weeks||72%||71%||0.99|
|Maintenance steroids (< 15 mg/day)||84%||67%||0.58|
|1-year survival rate following diagnosis||84%||93%||0.34|
|DAT-negative AIHA||DAT-positive AIHA||P-value|
|Efficacy of treatments||84%||80%||0.73|
|Treatment with steroids alone||70%||62%||0.99|
|Remission within 4 weeks||57%||43%||0.67|
|1 year survival rate following diagnosis||83%||80%||0.74|
|Treatment with steroids alone||n = 31||n = 13|
|Efficacy of steroids||83%||92%||0.65|
|Remission within 4 weeks||61%||50%||0.67|
|Maintenance steroids (< 15 mg/day)||85%||33%||0.14|
|1 year survival rate following diagnosis||83%||82%||0.74|
|Treatments||Idiopathic DAT-AIHA||Secondary DAT-AIHA|
|Splenectomy with steroids||2/2/2|
|Cyclophosphamide with steroids||4/3/3|
|Cyclosporine with steroids||1/0/0|
|Steroids+Cyclophosphamide+Intravenous gamma globulin||1/1/1|
|Intravenous gamma globulin||2/2/2|
|Intravenous gamma globulin with steroids||6/6/5|
|Danazol with steroids||1/1/1|
|Rituximab with steroids||2/1/1|
Based on multivariate analysis of idiopathic DAT−AIHA (n = 110), TP and Evans syndrome were identified as the only significant factors associated with the effectiveness of steroids (adjusted odds ratio [aOR], 1.36/[0.1 g/dl]; 95% CI, 1.01–1.84) and survival at the 1-year follow-up (aOR, 0.1; 95% CI, 0.01–0.88).
Our results indicate that patients with DAT−AIHA generally have milder anemia and hemolysis than patients with DAT+AIHA, respond equally well to steroids, and show comparable survival at 1-year follow-up. These data verify the empiric image of DAT−AIHA to show the same clinical conditions as DAT+AIHA, with the exception of an adequate amount of RBC-IgG, and also renew our awareness of the significance of diagnosis of C-AIHA, as several serious side effects of steroids make clinicians hesitant to use steroids to treat DAT-negative hemolytic anemia patients without a diagnosis of AIHA .
The similarity in the clinical characteristics between patients with DAT−AIHA and DAT+AIHA suggests that the clinical entity of DAT−AIHA might be nearly equivalent to DAT+AIHA, with the exception of RBC-IgG. This clinical consanguinity may discourage efforts to investigate the basic mechanism underlying DAT−AIHA, but it is assumed there is a difference between DAT−AIHA and DAT+AIHA, such as cytokine pattern  or IgG subclass . To elucidate the pathogenic mechanism of DAT-AIHA, one must have a thorough understanding of AIHA.
Multivariate analysis of patients with idiopathic DAT−AIHA revealed therapeutic and prognostic factors (TP and Evans syndrome, respectively) at the 1-year follow-up. In general, the natural history of AIHA cannot be predicted based on clinical features , but some early reports have suggested that a positive in DAT, reticulocytopenia or thrombocytopenia increased the mortality rate [23-25]. As these earlier reports were derived from data in the era of restricted available treatments, such as blood transfusion or splenectomy, whether or not thrombocytopenia affects prognosis using modern therapy is uncertain . Our data showed 10 deaths in patients with Evans syndrome, one-half of whom were treated with steroids and the causes of death were pneumonia (n = 2), pulmonary bleeding (n = 1), acute myocardial infarction (n = 1), and unknown (n = 1); the remaining patients were untreated or treated with blood transfusion alone and the causes of death were hepatic failure (n = 2), myelodysplastic syndrome (n = 1), sepsis (n = 1), and unknown (n = 1). Evans syndrome remains a risk factor. Recently, diabetes mellitus was reported to be associated with high early mortality and poor prognosis in patients with AIHA . We attempted to evaluate prognosis after intermediate or long intervals after diagnosis of DAT−AIHA with consideration for diabetes mellitus.
Our study had some limitations. First, because of the case–control study design, our data did not describe the prevalence or incidence of DAT−AIHA. Approximately 40% of undiagnosed hemolytic anemia referred to our laboratory was diagnosed with DAT−AIHA 1 year following the consult, which suggested that C-AIHA might be more common in AIHA than the rate previously assumed or that patients who were difficult to diagnose might merely accumulate in our laboratory. Secondly, the RBC-IgG value might affect the attending physicians in the diagnosis of DAT−AIHA more intensively than other clinical data. The RBC-IgG value was told to each attending physician within 1 week following the consult, because physicians struggled to diagnose their patients with unknown hemolytic anemia. The clinical diagnosis was assigned by attending physicians 1 year after the consult and 23% of untreated patients with high RBC-IgG values (>78.5 IgG/RBC) were diagnosed with non-AIHA because of other clinical data, and 28% of untreated patients with low RBC-IgG value (<78.5 IgG/RBC) were diagnosed with DAT−AIHA. Attending physicians appear to diagnose their patients using all their clinical data with respect to the RBC-IgG value.
In conclusion, our data characterized patients with DAT−AIHA in comparison with DAT+AIHA, verified the efficacy of steroid therapy in DAT−AIHA and identified significant factors for survival and effectiveness of steroids.
The authors sincerely thank Ms.Ohguri for her excellent technical assistance.
- 1Warm antibody autoimmune hemolytic anemia. Corticosteroid therapy. In: Petz LD, Garratty G, editors.Immune Hemolytic Anemias. Philadelphia:Churchill Livingstone;2004. pp401–407., .
- 12Clinical usefulness of specific antiglobulin reagents in autoimmune hemolytic anemias. Prog Hematol 1973;8:25–49..
- 13Clinically significant antibodies not detected by routine methods. In: Nance S, editor.Immune Destruction of Red Blood Cells. Arlington:American Association of Blood Banks;1989. pp93–99., .
- 14Unusual aspects of acquired immune hemolytic anemias. A autoimmune hemolytic anemia with a negative direct antiglobulin test (DAT). In: Petz LD, Garratty G, editors.Immune Hemolytic Anemias. Philadelphia:Churchill Livingstone;2004. pp319–334., .
- 17Immune destruction of red cells. In: Bell CA, editor.A Seminar on Immune-Mediated Cell Destruction. Washington, DC:American Association of Blood Banks;1981. pp93–130., , , et al.
- 19Acquired Immune Hemolytic Anemias, 1st ed. New York:Churchill Livingstone;1980., , editors.
- 21Coombs-positive and negative autoimmune hemolytic anemia might immunologically differ. Rinsho Ketsueki (Japanese J Clin Hematol) 2011;52:461., , .
- 22Prognosis and survival. Warm antibody AIHA. In: Petz LD, Garratty G, editors.Immune Hemolytic Anemias. Philadelphia:Churchill Livingstone;2004. pp70–71., .
- 26Diabetes mellitus is associated with high early-mortality and poor prognosis in patients with autoimmune hemolytic anemia. Hematology 2009;14:361–365., , , et al.