We read with interest Singer et al's report showing an increased alloimmunization and erythrocyte autoimmunization in a group of 64 transfusion-dependent thalassemia patients of predominantly Asian descent.1 These patients have been receiving phenotypically different red blood cells from predominantly white donors. We would like to share our experience in Hong Kong.
We have been treating a comparable number (68) of thalassemic patients (mean age, 18 years; range, 4-27 years) on regular transfusion. In this group 67 are of southern Chinese origin, and our blood donors are predominantly of the same ethnic origin. Of the 68 patients, 65 had β thalassemia major and 3 had HbE/β thalassemia heterozygocity syndrome. Using standard blood bank methods, serum was analyzed prior to each transfusion for detection of new antibodies to red blood cell antigens. We employed the same logarithm as Singer's for a further confirmation assay once antibody screening became positive. We always issue fully phenotypically matched blood to those who have an alloantibody in order to prevent further alloimmunization. We rarely need to transfuse K antigen–matched blood, contrary to what Singer et al had suggested.1 We had observed a total of 9 alloantibodies in 5 patients (of 68) with 7.4% detection rate (unpublished data, April 2001) as compared with 22% (14 of 64) as reported by Singer et al. Only 1 of our 68 patients had an autoantibody, compared with 6 of 64 in Singer et al's report.1 All our patients with alloantibodies have more than 15 years of regular blood transfusion (range, 15-26 years). Among the 9 alloantibodies detected in our patients (3 Anti-E; 3 Anti-Mi; 1 Anti-HLA; 1 Anti-BG, 1 Anti-BW22), only Anti-E has been reported to be clinically significant, causing a hemolytic transfusion reaction. Singer et al also noticed a relatively high rate of Anti-E (21%), and they could not attribute it to a recipient-donor antigenic discrepancy. Anti-K was not encountered in our patients, unlike the 6 of 19 (31%) anti-K alloantibodies in Singer et al's group. No single alloantibody for c, S, and Fyb was found in our patients. All of our patients' alloantibodies were developed before we introduced universal leukodepletion for thalassemic patients. This observation is in line with Singer et al's observation that a significantly lower alloantibody rate had resulted from the introduction of leukodepletion.1 Twelve our patients underwent splenectomy, but only 1 of those 12 had alloantibodies and she developed the antibody before splenectomy. And only 1 of the 12 had autoantibodies after splenectomy. Our experience did not substantiate Singer et al's observation that patients who underwent splenectomy had a higher alloimmunization rate.1
Previous data on a presumed homogenous population in Greece and Italy also showed an overall low rate (5% to 10%) of alloimmunization.2,3 The difference between our experience of lower rate of alloimmunization (7.4%) and Singer et al's higher rate (22%) can be explained by our access to phenotypically matched donors in Hong Kong. We agree with Singer et al's recommendation that the recruitment of Asian blood donors in North America, just like the recruitment of black donors for sickle cell disease patients, can increase the availability of compatible blood for thalassemia patients, who have a lifelong need for transfusions.
Differences in alloantibody types detected in transfused Asian thalassemia patients
Ho et al add important data for better understanding the causes for alloimmunization in Asian thalassemia patients undergoing transfusion. Their data assists in delineating these causes among homogenous and less homogenous donor-recipient populations. The differences in the type of antibodies detected in our and Ho et al's study may also be largely due to the red blood cell phenotypic discrepancy.
We agree with Ho et al that the effect of absence of spleen on allo- and autoantibody development is not completely clear. Our data suggested a higher frequency among patients in which the spleen was removed but has not reached statistical significance. Larger studies, in patients from various ethnic backgrounds, are needed to address this issue.