0 1984 S. Karger AG. Basel 0042-Y007/84/0473-0205 $2.75/0 Vox Sang. 47: 205-208 (1984)

Should Chronic Transfusions Be Matched for Antigens Other than ABO and Rh,(D)?

Neil Blumberg, Karen Ross, Eduardo A d a , Kathy Peck Blood Bank, Clinical Pathology Laboratories and Department of Pathology, University of Rochester Medical Center, Rochester, N.Y., USA

Abstract. It has been recommended that red blood cell transfusions to patients with hemoglobinopathy or aplastic anemia be matched for antigens other than ABO and Rh,(D). We studied 1,010 patients with disorders that often lead to repetitive transfusion. The fre- quency of transfused patients with clinically important antibodies was not significantly different among the disease groups except for those with lymphocytic leukemia. The fre- quency ofmultiple red cell antibodies was about 3% overall. Most antibodies (71%) devel- oped early in the transfusion course, before the 15th transfusion. From the standpoints of frequency of alloimmunization, multiplicity of antibodies, and time course of antibody development, patients with hemoglobinopathy and aplastic anemia were not significantly different from other transfused patients. Matching for antigens other than ABO or Rh,(D) might increase costs in our hospital by 40,000-370,000 dollars per year for these patients. Because morbidity or mortality due to these antibodies is rare, antigen matching for other than ABO and Rh,(D) is not cost-effective.

Introduction

One hazard of blood transfusion is the formation of alloantibodies that complicate the provision of further safe transfusion. Transfusions routinely are matched for ABO and Rh,(D). However, it has been ad- vocated that patients with hemoglobinopa- thies and with aplastic anemia should re- ceive transfusions matched for ‘minor’ blood group antigens, such as those in the Rh, Kell, Duffy, Kidd and Lewis systems, in an at-

tempt to prevent alloimmunization [I-31. Several studies have examined the frequency of alloimmunization in these populations [4-81. However, no study has evaluated the cost-effectiveness of matching for antigens other than ABO or Rh,(D). We retrospec- tively studied the frequency of alloimmuni- zation in 1,010 transfused patients with gas- trointestinal bleeding, renal failure, hemo- globinopathy, aplastic anemia, or leukemia and performed a cost analysis for antigen matching.

206 Blumberg/Ross/Avila/Peck

Methods

The method of patient data acquisition was similar to that in our previous study [8]. All patients with dis- charge diagnoses of leukemia, aplastic anemia, gastro- intestinal bleeding, chronic renal failure, and hemoglo- binopathy during 1970-1980 at our institution were identified by searching a computer data base. Due to the large number of patients with diagnoses of gastrointes- tinal bleeding or chronic renal failure, only those pa- tients discharged between 1978 and 1980 with these diagnoses were included. Data are reported for all patients who received transfusions. Alloimmunization only to clinically significant antigens was considered. Statistical analysis was performed as we have pre- viously reported [8,9]. Detailed descriptions of the cost estimates for antigen matching are available from the authors. The estimated incremental costs in our blood transfusion service were $140 per unit ofblood matched for all four minor antigen systems (Rh, Kell, Kidd, Duffy), $30 per unit for three systems (Rh, Kell, Kidd, or Duffy), and $15 per unit for Rh and Kell alone.

The mean number of transfusions, the total number of transfused patients, and the

frequency of patients with red blood cell antibodies and of patients with multiple an- tibodies are shown in table I for each disease group. There were no significant differences in the frequency of patients with antibodies except that patients with lymphocytic leu- kemia had fewer antibodies detected than those with gastrointestinal bleeding (p <0.001), renal failure (p < 0.005), he- moglobinopathy (p< myeloid leu- kemia (p < 0.0005), and aplastic anemia (p<0.005).

The distribution of blood group speci- ficities of clinical significance was similar in all disease groups. Antibodies in the Rh and Kell groups accounted for 87% of the 126 antibodies generally considered clinically significant. The individual Rh specificities were similar in each disease group.

Figure 1 displays when antibodies were initially detected in the course of transfu- sion. More than a quarter (27%) of the anti- bodies were present on initial testing, and almost three-quarters (71%) were detected

Table I. Frequency of patients with red blood cell antibodies in the disease groups

Group Total number Number of trans- Patients with Patients with of patients fusions antibody(-ies) multiple

antibodies

mean range n TO n Oh

Gastrointestinal bleeding 237 22 1-256 18 8 7 3 .O Renal failure 20 I 49 3-414 19 10 9 4.5 Hemoglobinopath y 89 33 1-184 15 17 4 4.5 Myeloid leukemia 209 19 1-122 18 9 6 2.9 Lymphocytic leukemia 193 10 1-115 1 I 0 0 Aplastic anemia 81 33 1-299 9 1 1 5 6.2

Total 1,010 26 80 7.9 31 3.1

Antigen-Matched Transfusions 207

n

I I I I I 0 1-4 5-14 15-24 25-34 35-44 45-54 55-64 ?65 Number of transfusions

Fig. 1. Relationship between the initial appearance of an antibody and number of transfusions.

before the 15th transfusion. The rate ofanti- body formation per 1,000 transfusions was 7.9 (transfusions 1-14), 4.3 (15-44) and 2.5 (1 45) and these rates were significantly dif- ferent from each other. The time course of antibody development was not significantly different in the various disease groups. The sex ratio of males: females was 1 : 2 for those patients with antibodies detected before their first transfusion in our institution. The

sex ratio was, however, 2 : 1 for those making antibodies after their first transfusion.

Table I1 shows the additional cost of test- ing in 1981 dollars had the transfusions shown in table I been matched for two, three or four blood group systems in addition to the usual ABO and Rh,(D).

Discussion

Despite great differences in numbers of transfusions and disease process, patients requiring transfusion for the conditions studied vary little, if at all, in their tendency to have single or multiple red blood cell anti- bodies, or in time of appearance of these antibodies. However, as shown previously [8], transfused patients with lymphocytic leukemia had significantly less likelihood of having blood group antibodies. There seems little basis for choosing patients with hemo- globinopathy or aplastic anemia to receive more completely antigen-matched blood

Table 11. Added costs for matching transfusions given for additional blood group systems'

Disease group Number of Rh and Kell Rh, Kell, and Rh, Kell transfusions dollars Duffy or Kidd Duffy and Kidd

( 1 5/unit) dollars dollars (301unit) ( 1 40/unit)

Gastrointestinal bleeding 5,214 Renal failure 9,849 Hemoglobinopath y 2,937 Myeloid leukemia 3,97 I Lymphocytic leukemia 1,930 Aplastic anemia 2,613

~

Total (10 years) 26,574

78,000 156,000 730,000 148,000 296,000 1,379,000 44,000 88,000 4 1 1,000 60,000 120,000 556,000

29,000 58,000 270,000 40,000 80,000 374,000

399,000 798,000 3,720,000

I 198 1 dollars (rounded to nearest $1,000).

208 Blumberg/Ross/Avila/Peck

than patients with conditions entailing sim- ilar longevity and chronicity of transfusion (e.g., some patients with chronic renal fail- ure or gastrointestinal bleeding).

The data in figure1 suggest that most blood group antibodies seen in multiply transfused patients are due to previous preg- nancy and to the initial 1-10 transfusions. The rate of antibody formation per transfu- sion actually decreases with increasing num- bers of transfusions. These findings argue against the notion that alloimmunization is primarily a phenomenon of chronic transfu- sion. Rather, they suggest that many individ- uals capable of making a blood group anti- body make that antibody during the first few exposures to the appropriate antigen.

The information in table I1 suggests that even modest (e.g., Rh and Kell) matching of transfusions for additional antigens, even for selected patients, could increase costs dra- matically. To provide transfusions matched for more antigens to any significant number of repetitively transfused patients would be prohibitively expensive. While the question of transfusions matched for multiple anti- gens is a financial question, it is also one of medical priorities and logistics. Only 3% of the patients studied developed more than one antibody. None of the patients in this study had a clinically detectable hemolytic reaction. Extensive antigen matching leads to needless delays in transfusion.

Our current recommendations are that patients should generally receive red cells matched only for Rh,(D) and ABO.

Acknowledgements

We are indebted for helpful discussions to Drs. D. Arvan, C. Chuang, H. Cohen. J . Heal, J . Nusbacher,

E. Snyder, and Ms. A. McMican. Mrs. Rebecca Cadre- gar; and Ms. Carol Vun Voorhees provided excellent secretarial assistance.

References

Denes, A. E.: Anemia, in Freitag, Miller, Manual of medical therapeutics; 23rd ed., p. 279 (Little, Brown, Boston 1980). Cohen, H.J.; Lipton, J.M.: Blood disorders, in Graef, Cone, Manual of pediatric therapeutics; 2nd ed., p.428 (Little, Brown, Boston 1980). Kevy, S. V.; Jacobson, M.; Button, L.: Clinical uses of frozen-thawed erythrocytes in pediatrics. Prog. din. biol. Res. 1I: 89-95 (1976). Economidou, J. ; Constantoulakis, M. ; Augoustaki, 0.; Adinolfi, M.: Frequency of antibodies to van- ous antigenic determinants in polytransfused pa- tients with homozygous thalassaemia in Greece. Vox Sang. 20: 252-258 (1971). Blumberg, B.S.; Alter, H. J.; Riddell, N.M.; Er- landson, M.: Multiple antigenic specificities of se- rum lipoproteins detected with sera of transfused patients. Vox Sang. 9: 128-145 (1964). Orlina, A. R.; Unger, P. J.; Koshy, M.: Post-trans- fusion alloimmunization in patients with sickle cell disease. Am. J. Hematol. 5: 101-106 (1978). Coles, S. M.; Klein, H. G . ; Holland, P. V.: Alloim- munization in two multitransfused patient popula- tions. Transfusion 21: 462-466 (1981). Blumberg, N.; Peck, K.; Ross, K.; Avila, E.: Im- mune response to chronic red blood cell transfu- sion. Vox Sang. 44: 212-217 (1983). Snedcor, G. W.; Cochran, W.G.: Statistical meth- ods, pp. 115-116,124-128,166,208-213 (IowaState University Press, Ames 1980).

Received: May 5, 1983 Accepted: October 10, 1983

Dr. Neil Blumberg, Box 608, University of Rochester Medical Center, Rochester, NY 14642 (USA)