Autoimmune hemolytic anemia (AIHA) is a rare, potentially fatal condition that often requires prolonged immunosuppressive therapy. In the West, its incidence is 1–3 per 10⁵ per year and its prevalence is 17:100,000 [1]. There is a lack of data about the incidence of AIHA in India. This disease is often not recognized and diagnosis is delayed. Autoantibodies produced against red cell antigens cause hemolytic anemia in this condition. This may be a primary (idiopathic) process or a secondary process caused by other diseases such as autoimmune disorders, lymphoproliferative disorders, infections, or tumors [23]. Presentation is varied, ranging from insidious onset of anemia with hyperbilirubinemia progressing over months to acute fulminant hemolysis occurring within hours to days and leading to a sudden drop in hemoglobin. The diagnosis is made using the direct antiglobulin test (DAT) with polyspecific antiglobulin reagents. Two serological types of the disease, warm autoimmune hemolytic anemia and cold agglutinin disease, are recognized based on the thermal range of the autoantibodies. In warm autoimmune hemolytic anemia (WAIHA), the DAT usually yields positive results with anti-immunoglobulin G (anti-IgG) antisera and, in CAD, the DAT yields positive results with anti-C3d antisera and indicates the presence of high titer cold agglutinins. The autoantibodies produced in CAD usually belong to the IgM subclass. In some proportion of cases, the thermal range of antibodies may be substantially wide enough to cause hemolysis in both warm and cold conditions. These cases are detected via the coexistence of warm antibodies and high titer cold agglutinins and are designated as mixed type AIHA. Even with the availability of sensitive techniques such as the microalbumin test, flow cytometry, and mitogen-stimulated DAT [45], in one-tenth of AIHA cases, the DAT can demonstrate negative results, and these cases are designated as atypical AIHA [1]. Because of a lack of prospective clinical trials [6], the treatment of AIHA is based on large clinical studies [789], experts' opinions, and individual experience. While the first option for treatment of WAIHA is usually corticosteroids, there is no consensus for the treatment of CAD. The choice of second-line agents in cases of relapsed disease is based on a physician's comfort with various immunosuppressive agents such as azathioprine, cyclosporin, mycophenolate mofetil (MMF), and anti-CD20 monoclonal antibody (rituximab) and splenectomy. There is a lack of data concerning the diagnosis and treatment of AIHA in India. With this background, we decided to study the clinical profile of this disease, its response to frontline and second-line agents, and the occurrence of complications and to correlate this information with the different serological types of the disease.

Because AIHA is a rare disease with varied presentation, serological characteristics, and severity, it is difficult to conduct properly designed prospective studies. As it is a benign disease and thus less 'glamorous' than other hematological disorders such as leukemia and myeloma, hemato/oncologists often disregard this potentially fatal disease. Because of a lack of well-established guidelines and heterogeneity, therapy is a challenge and often prolonged therapy is required. There is dearth of data from South India. We retrospectively analyzed our patients. Given the rarity of the disease, our numbers are considerable, especially as this was a single center study. Similar to a previous, large GIMEMA study [1], we had a predominantly female patient population. However, in contrast to that study, the median age of our population was lower (41 yr) and the majority belonged to the young adult group (18–40 yr). This younger age of onset in the Indian population is in agreement with a previous observation from northern India where the median age was 31 years [10]. However, that study included both primary and secondary AIHA patients. In contrast to previous studies [111213], three-fourth of our patients had severe anemia at presentation with about 42% having reticulocytopenia, probably due to immune-related destruction of reticulocytes and apoptosis of red cell precursors [14]. One interesting observation was that younger patients had a severe onset of disease. The reason for this could be diagnostic delay, as nutritional anemia is the commonest cause of anemia and patients with anemia are often initially treated with hematinics before they are referred to specialists for evaluation. Upfront systematic evaluation of anemia is important as earlier initiation of therapy for AIHA lowers the probability of relapse [23].

We had an equal proportion of WAIHA and CAD patients. This is in contrast to previous reports that had a preponderance of warm cases [23]. Only 6% of our cases were DAT negative atypical AIHA. This is in contrast to most previously reported series, but in agreement with a recent large observational study [1]. We used a more sensitive gel technique rather than conventional tube tests, which explains the low percentage of atypical AIHA. Surprisingly, in these atypical cases, the average Hb was higher than DAT positive cases. One expects a certain diagnostic delay in DAT negative cases and, consequently, a drop in Hb while awaiting the completion of diagnostic work. One possible explanation is that such cases produce a much lower level of autoantibodies than the detection threshold of current techniques and, consequently, the rate of hemolysis is indolent and leads to a slow decline in Hb. In our series, the average Hb was comparable between WAIHA and CAD cases. One expects greater hemolysis and, consequently, more severe anemia in CAD due to IgM-mediated intravascular hemolysis than in WAIHA, which involves IgG-mediated extravascular hemolysis [23].

All of the 30 patients who required treatment were treated with upfront high-dose corticosteroid therapy. Among them, 90% achieved a response irrespective of the serological type. More patients with CAD achieved CR than WAIHA patients, though this was not statistically significant. This contrasts with most previous reports, but is in agreement with a previous report from India [10] where the overall response to corticosteroids was 88% regardless of serological type. Although a large number of patients respond to initial corticosteroid therapy, it is well-established that less than a fifth are cured by corticosteroids alone and maintain persistent complete remission [2315]. Our study agrees with this observation; 13% of the patients who required steroid therapy are in CR and off medications and another 17% are in CR with a low dose of steroids (<15 mg/d prednisolone). The median duration of steroid treatment in this series was 14 months. It has been shown that patients who continue to receive low dose corticosteroids for more than 6 months have a lower chance of relapse than those in whom corticosteroids were discontinued before 6 months [16]. Although recent recommendations advise against corticosteroid therapy for CAD [9], we found a good response with corticosteroids in these patients. It should be noted that we did not differentiate between mixed type AIHA, where antibodies have a large thermal amplitude, and CAD, which, to some extent, may explain the better response seen in CAD cases. However, according to most previous reports, one only expects around 10% of mixed cases and this cannot fully offset the good results we observed in our CAD cases [1].

The small difference between median steroid duration (14 mo) and median TTNT (19 mo) indicates that the majority of patients who are off corticosteroids will eventually relapse and require further agents. The assessment of the efficacy of second-line agents is blurred by the fact that patients are simultaneously on second-line agents and corticosteroids, and it is often not possible to exactly delineate the benefit of the second-line agent. We have had maximum experience with azathioprine in relapsed cases and, in contrast to the findings of a recent large study [1], nearly four-fifths of our patients responded durably regardless of the AIHA type. We have had limited experience with other immunosuppressive agents and hence cannot draw any inference of their effect from this study. Of note, we did not see a response in both patients who received MMF. This is an interesting observation, especially as MMF has been proven to be effective in post-hematopoietic stem cell transplant AIHA [1718]. The extensive use of azathioprine instead of rituximab, which is recommended as a first-line agent for CAD [9], is because rituximab is unaffordable to most in a country where patients pay out of their pocket.

Of the 15 patients who required therapeutic agents beyond steroids, 4 received rituximab therapy; 1 received it as a second-line agent and 3 as a third-line agent. All of these patients achieved at least PR. A response was also observed in patients administered rituximab LD. This is interesting as rituximab LD would substantially save costs. However, according to recent reports, larger prospective studies of rituximab LD are needed before making it a routine clinical practice, especially in CAD cases [61920].

Interestingly, we did not treat a single patient with splenectomy. This possibly indicates the availability of more therapeutic agents and the reluctance of patients to undergo a surgical procedure that has a low but significant risk of post-splenectomy sepsis. In line with the severe presentation, two-thirds of our patients initially required PRBC transfusions. However, all of them were transfusion-independent one week into steroid therapy.

Our series of primary AIHA cases shows that nearly half of our patients required two lines of therapy and one-fifth required a third agent. A younger age was associated with a more severe clinical presentation. There was no difference in response to corticosteroids among WAIHA and CAD cases. In relapsed cases, four-fifths of the cases responded to azathioprine in conjunction with corticosteroids and half of these patients did not require a third agent.

Two patients in our series had venous thrombosis. Of these, 1 patient had a stormy clinical course with pneumonia and required 3 lines of therapy. Both of them did not have anticardiolipin antibodies. An association has been described in AIHA between venous thrombosis and anticardiolipin antibodies [21]. Interestingly, both patients who developed venous thrombosis had ES, which indicates that patients with ES are at a higher risk for thrombotic events than those with AIHA alone. Prothrombotic risk in ES cases has been recognized before [22]. There were 4 infective episodes: 1 was grade 2, 2 were grade 4, and 1 was grade 5. The patient who died had concomitant chronic liver disease and bacteremia and developed severe sepsis.

In conclusion, our study is the first from South India to observe the clinicoserological characteristics of AIHA patients. This is probably the first study from India to examine the effectiveness of second-line agents in AIHA. We found that a younger age of onset is associated with severe disease presentation. We observed an equal proportion of warm and cold cases. We found that contrary to recent reports, cold cases responded well to upfront corticosteroids. Although an observational study is not ideal for making treatment recommendations, we found that azathioprine is an effective second-line agent. With the availability of rituximab and because of the risk of severe post-splenectomy sepsis, the role of splenectomy has considerably decreased. Finally, AIHA is a very heterogeneous disease with a potential for fatal outcome that requires care by physicians experienced in treating these patients.