Lack of Predictive Value of Antileukocyte Antibody Screening in Granulocyte Transfusion Therapy R. s. UNGERLEIDER, F. R. APPELBAUM, R. J. TRAPAN1 AND A.
€3. DEISSEROTH
From the Experimental Hematology Section, Pediatric Oncology Branch, National Cancer Institute.
Bethesda, Maryland
To clarifj the relationship between recipient presensitizstion and response to granulocyte (PMN) transfusion, we tested 187 non-HL-A matched donor-recipient pairs for the presence of antikukocyte antibody using granuhxytotoxiclty (G), lymphocytotoxicity (L), microkukoagglutination(MI, and capillaryleukoegglutination (C)assays. PMN increments per 10" transfused PMNs per square meter of body surface area, ascertained one hour following termination of transfusion, and the occurrence of nonhemolytic transfusion reactions, were correlated with the assay results. Although circulating anti-donor-leukocyte antibody was detected in 52 per cent of recipients, there was no statistically significant relationship between the presence of these antibodies and either PMN recovery or incidence of transfusion reaction. We conclude that the prospective use of these assays is of little value in predicting the recipient's response to PMN transfusion.
GRANULOCYTE (PMN)transfusion therapy is of recognized clinical benefit in infected neutropenic patient^,'*^*^ and may be lifesaving in persistently granulocytopenic indiv i d u a l ~It. ~has been suggested that survival of PMNs in the transfused host is compromised by the presence of preformed alloantibodies directed against donor cells, as evidenced by decreased circulation of PMNs and by nonhemolytic transfusion reactions.e-8*'oAssays for the detection of antileukocyte antibodies in the form of leukoagglutinins, lymphocytotoxins and granulocytotoxins have been devised with the intent of providing information useful in selecting donors and avoiding transfusion reactions.* We have employed four such assays in an attempt to clarify the relationship between recipient presensitization and response to PMN transfusion. Supported in part by NIH contract RNOI-CM63810. Received for publication December 23. 1977; accepted February 20, 1978.
Materials and Methods One hundred and eighty-seven PMN transfusions given to 19 neutropenic cancer patients during 20 episodes of septicemia were evaluated. All patients had been entered in a randomized study of the utility of PMN transfusion therapy, and had been multiply transfused prior to the septic episode. PMNs were collected using the continuous flow centrifuge (CFC) in all instances. All transfusions were ABO compatible but were not HL-A matched. The median number of transfusions per recipient was six. Recipients were pretreated with diphenhydramine, hydrocortisone, acetominophen, and or chlorpromazine in 15 instances. Recipient sera were obtained twice weekly during the period of transfusion, stored at -70 C, and examined daily for the presence of antileukocyte antibody against the donor. The techniques for these assays, published previously,2 are briefly outlined below. Granulocytotoxicity ( G )
Serial twofold dilutions of heat-inactivated recipient serum were mixed with a standardized donor PMN suspension. Following a 3lXminute incubation at 23 C, complement was added and incubation continued for another 60 minutes. Eosin Y and formaldehyde were then added, the cells were allowed to settle, and the degree of granulocytotoxicity was ascertained employing an inverted phase contrast microscope. Controls .included a negative serum (the serum autologous to the PMNs), a known positive serum, and a buffer control. Lymphocytotoxicity ( L )
Serial twofold dilutions of heat-inactivated recipient serum were mixed with a standardized donor lymphocyte suspension collected at the lymphocyte-rich interface layer of FicollHypaque gradients. Following incubation at 23 C for 60 minutes, complement was added and lymphocytotoxicity determined in a manner identical to that outlined for granulocytotoxicity.
0041-1132/79/0100/0090$00.75 0 J. B. Lippincott CO. Transfusion January-February 1979
90
Volume 19 Number I
Volume 19 Number I
.
91
ANTILEUKOCYTE ANTIBODY SCREENING
Controls consisted of the serum autologous to the lymphocytes, a buffer control, and a known positive serum.
Table 1. Incidence of Positive Antileukocyte Antibody Assays
Assay
No. of Positive Samples
Per cent of Total Samples (n = 187)
Granulocytotoxicity Lymphocytotoxicity Microleukoagglutination Capillary leukoagglutination
5 6 9 87
3 3 5 47
MicroleukoaK~Iutination( M )
Serial twofold dilutions of heat-inactivated serum were mixed with a standardized donor leukocyte suspension and incubated at 37 C for two hours. Toluidene blue was then added and the mixtures were examined using an inverted light microscope. The degree of leukocyte agglutination was scored on a 0 (no agglutination) to 4+ (strong agglutination) scale. Controls were run as above. Capillary Leukoagglutination (C)
Serial twofold dilutions of heat-inactivated recipient serum were mixed with a standardized donor leukocyte suspension. Following I5 minutes of incubation at room temperature, capillary tubes were filled with aliquots of the reaction mixture, sealed at one end, and centrifuged. The capillary tubes were then maintained at a 45 degree angle with the leukocyte pellet uppermost, for 60 minutes at 23 C. The length of leukocyte migration (mm) was measured employing a phase contrast microscope. Each assay was run in tandem with both positive and negative control sera. Assays were set up in triplicate and the mean migration length of the leukocytes was employed for calculation of the per cent inhibition of leukocyte migration in comparison to migration in autologous (negative) serum. One hour posttransfusion PMN increments per l o l l transfused PMNs per square meter of body surface area (corrected 1 hour PMN increments) were calculated and compared with the assay results. The response to transfusion was not known to the individual performing the in vitro assays. Assays were considered positive if, at any dilution up to 1:512, there was greater than 10 per cent cell death or 1 + agglutination; or if there was 10 per cent inhibition of migration up to a dilution of 1: 16. Records of each PMN transfusion administered to these patients were examined for correlation between presence of antibody and occurrence of nonhemolytic transfusion reactions (chills, fever, dyspnea, cyanosis, and/or hypotension). At the time of transfusion, a nurse had been in attendance monitoring and recording vital signs every ten minutes. A physician was called to evaluate the patient in the event of suspected reaction, and the clinical impression recorded on a protocol sheet.
Results Ninety-six of the 187 transfusions (52%) were associated with the presence of antibodies against donor leukocytes as determined by one or more of the assays used. The majority of these antibodies was detected by the C assay (Table 1). Ten donor-recipient pairs (5%) had antibody demonstrated by more than one assay. There were six instances (3%) in which the C assay was negative while antibody was detected by one of the other methods. The mean corrected 1 hour granulocyte increment for all patients was 2149 (SD = 9073), with a range of -43,683 to +54,954. The extreme high and low increments were seen following transfusion in which small doses of PMNs were given to large patients and the increments then corrected to 10" transfused cells per square meter body surface area. Transfusions in which the recipient serum demonstrated no preformed antibody produced a mean corrected 1 hour PMN increment of 2803 (SD = 6637) while presence of antibody was associated with a mean increment of 1487 (SD = 11,OOO). Although the increment appeared to be greater in the group without preformed antibody, the within-group variance was extremely large (Fig. l), and the difference between the two groups was not statistically significant (p > 0.10; Wilcoxon Rank Sum Test). One hundred twenty-five transfusions resulted in PMN increments at 1 hour. The influence of preformed antileukocyte antibody on PMN recovery is presented in Table 2. Chi-square analysis revealed no statistically significant relationship between the presence of antibody and PMN recovery ( x 2 = 0.007, p > 0.10). Particular attention was given to the capillary leukoagglutination assay, as this technique most frequently detected antibody. In the presence of circulating leukoagglutinins as detected by this assay, 64 per cent of recipients achieved granulocyte increments compared to 66 per cent of leukoagglutinin-free recipients. This difference
92
UNGERLEIDER ET AL.
56,000
r
Transfusion January-February 1979
0
I-
FIG.1. Distribution of corrected granulocyte increments as a function of antileukocyte antibody assay results.
0
a 3
9 z"
0
-8,OOo
n
za -16,000 a
8 -24,000
-32,0004,OOo
-
1
I
was not statistically significant ( x ' p > 0.10).
I
= 0.0037,
Nonhemolytic transfusion reactions occurred in 9/187 (5%) of these transfusions. While none of the reactions was considered lifethreatening, three of them were characterized by respiratory compromise reflected in cyanosis of
I
I
the nailbeds and/or dyspnea, and were considered severe. As shown in Table 3, the presence of preformed circulating antileukocyte antibody detected by any of our assays did not influence the incidence of reaction ( x 2 = 0.75, p > 0.10). Separate analysis of the C assay (x' = 0.30, p > 0.10) and the G, L and M assays
Volume 19 Number 1
93
ANTILEUKOCYTE ANTIBODY SCREENING
Table 2. Influence of Preformed Antileukocyte Antibody on 1 Hour Granulocyte Recovery.
Antibody
No. Transfusions Producing Increments
No. Transfusions Producing No Change or Decrements
Present Absent
61 (33%) 64 (33%)
29 (16%) 33 (1 So/,)
( x 2 = 1.37,
p > 0.10) uncovered no statistically significant correlation. On two occasions, a recipient with a positive C assay was premedicated with hydrocortisone and diphenhydramine, and may have had a potential reaction aborted by these drugs. Adding these two patients to the group having a positive C assay with reaction, and subtracting them from the group having a positive C assay without reaction, did not reveal a statistically significant correlation ( x 2 = 1.37, p > 0.10). The distribution of nonhemolytic transfusion reactions as a function of levels of antibody as detected by the C assay is presented in Table 4. Chi-square analysis indicated that the relationship between per cent inhibition of migration of leukocytes and severity of reaction did not differ from chance (xz = 3.18, p > 0.10). This analysis was not extended to G, L, or M assays as only two reacting recipients had positive results .
Discussion
We have investigated the relationship between the presence of preformed antileukocyte antibody and the response of the recipient to transfused PMNs. Using a battery of four immunohematologic assays, we detected antibody against donor leukocytes in the majority of our recipients. We were unable, however, to demonstrate a significant relationship between the presence of such antibodies and either decreased recovery of circulating PMNs or increased incidence of nonhemolytic transfusion reactions. Thus, our findings indicate that the prospective use of these assays in screening for antileukocyte antibody is not of value in predicting successful donor-recipient pairs in terms of reaction-free transfusions or posttransfusion increments. It is admittedly difficult to evaluate the
Table 3. Influence of Preformed Antileukocyte Antibody on Incidence of Nonhemolytic Transfusion Reaction
No. of Transfusions (n = 187) Antibody Assay
Assay Result
With Reactions
Without Reactions
C, G, Lor M
Positive Negative
6 (3%) 3 (2%)
82 (44%) 96 (51%)
C
Positive Negative
5 (3%) 4 (2%)
72 (39%) 106 (56%)
G, L or M
Positive Negative
2 (1%) 7 (4%)
11 (6%) 167 (89%)
response to an individual granulocyte transfusion. We have used the occurrence of transfusion reactions and the 1 hour posttransfusion increment because they are clearly definable parameters and are clinically relevant. Reaction-free transfusions are obviously desirable, and the circulating granulocyte level is known to correlate with the incidence and severity of infectious episode^.^ Prior studies have suggested that presensitization is of consequence in patients receiving PMN replacement therapy. Wellcontrolled animal studies employing these assays have shown that prior alloimmunization results in decreased circulation of granulocytes.2 In addition, transfusion of leukoagglutinins has been associated with severe pulmonary reaction in humans.'O Although the results of the present study do not support these findings, there are mulTable 4. Effect of Per Cent Inhibition of Leukocyte Migration as Measured by C Assay on Incidence of Nonhemolytic Transfusion Reaction' Per cent Inhibition of Leukocyte Migration
0-9 10-30 40-60 70- 100
No. of Transfusions Associated with
No Reaction
Mild Reaction
Severe Reaction
99
3
47
2 1 0
2 1
19
13
0 0
UNGERLEIDER ET AL.
tiple factors which may obscure the interpretation of response to granulocyte transfusion in the usual clinical setting. Recipients are found to be heterogeneous, with different degrees of neutropenia and hypersplenism, which may affect circulation of granulocytes. Concurrent drug therapy may alter the patient’s ability to mount an immune response, or, as with steroids, change the rate of entry and egress of granulocytes from the b l o ~ d .In~ addition, there are technical limitations imposed when counting low levels of leukocytes, which are then compounded by similar limitations in doing differentials on leukopenic individuals. Small errors of measurement under these circumstances become magnified when calculations are performed to account for differences in body surface area. It becomes apparent, then, that the wide variations in recorded increments could mask the consequences of alloimmunization when increments are used as the criterion. Similarly, recognition of nonhemolytic transfusion reactions can be affected by differences in drug therapy, mental status, and previous blood component therapy, as well as awareness of the observer. It is possible that with larger doses of infused granulocytes, more sensitive techniques of measuring granulocyte circulation, and improved methods of detecting antibody, pretransfusion screening will be of importance. At the present time, however, given the limitations outlined above, the prospective use of these assays is of little clinical utility. References 1. Alavi, J. B., R. K. Root, I. Djerassi. A. E.
Evans, S. J. Gluckman, R. R. MacGregor, D. Guerry, A. D. Schreiber, J. M. Shaw, P. Koch, and R. A. Cooper: A randomized trial of granulocyte transfusions for infection in acute leukemia. N. Engl. J. Med. 296:706, 1977. 2. Appelbaum, F. R., R. J. Trapani, and R. G. Graw,
Trmsfhsion JMWFebruary 1979
Jr.: Consequences of prior alloimmunization during granulocyte transfusion. Transfusion 17: 460, 1977. 3. Bishop, C. R., J. W. Athens, D. R. Boggs, H. R.
Warner, G. E. Cartwright, and M. M.Wintrobe: Leukokinetic studies. XIII. A non-steady-state kinetic evaluation of the mechanism of cortisone-induced granulocytosis. J. Clin. Invest. 47: 249, 1968. 4. Bodey, G. P.,
5. 6.
7.
8.
M. Buckley, Y. S. Sathe, and E. J. Freireich: Quantitative relationships between circulating leukocytes and infection in patients with acute leukemia. Ann. Intern. Med. 64:328, 1966. Boggs, D. R.: Neutrophils in the blood bank. N. Engl. J. Med. 2%:748, 1977. Cooper, M. R., E. Heise, F. Richards, J. Kaufmann. and C. L. Spurn: A prospective study of histocompatible leukocyte and platelet transfusions during chemotherapeutic induction of acute myeloblastic leukemia. I n : Leukocytes: Separation, Collection and Transfusion. J. M. Goldman and A. M. Lowenthal, Eds. London, Academic Press, 1975, p. 436. Goldstein, I. M., H. J. Eyre, P. I. Terasaki, E. S. Henderson, and R. G. Graw, Jr.: Leukocyte transfusions: Role of leukocyte alloantibodies in determining transfusion response. Transfusion 11:19, 1971. Graw, R. G. Jr., G. Herzig, S. Perry, and E. S. Henderson: Normal granulocyte transfusion therapy. Treatment of septicemia due to gramnegative bacteria. N. Engl. J. Med. 282367,
1972. 9. Herzig, R. H., G. P. Herzig, R. G. Graw, Jr..
M. I. Bull, and K. K. Ray: Successful granulocyte transfusion therapy for gram-negative septicemia. N. End. J. Med. 2%:701, 1977. 10. Thompson, J. S., C. D. Severson, M.J. Parmely, B. L. Marmorstein, and A. Simmons: Pulmonary “hypersensitivity” reactions induced by transfusion of non-HL-A leukoagglutinins. N. Engl. J. Med. 284:1120, 1971.
Richard S. Ungerleider, M.D.,Investigator, Pediatric Oncology Branch, National Cancer Institute, Bethesda, MD 20014. Frederick R. Appelbaum, M.D., Investigator, Pediatric Oncology Branch. Robert-John Trapani, Director, Department of Immunology, Microbiological Associates, Bethesda, MD 20014.
Albert B. Deisseroth, M.D., Ph.D., Head, Experimental Hematology Section, Pediatric Oncology Branch.
€3. DEISSEROTH
From the Experimental Hematology Section, Pediatric Oncology Branch, National Cancer Institute.
Bethesda, Maryland
To clarifj the relationship between recipient presensitizstion and response to granulocyte (PMN) transfusion, we tested 187 non-HL-A matched donor-recipient pairs for the presence of antikukocyte antibody using granuhxytotoxiclty (G), lymphocytotoxicity (L), microkukoagglutination(MI, and capillaryleukoegglutination (C)assays. PMN increments per 10" transfused PMNs per square meter of body surface area, ascertained one hour following termination of transfusion, and the occurrence of nonhemolytic transfusion reactions, were correlated with the assay results. Although circulating anti-donor-leukocyte antibody was detected in 52 per cent of recipients, there was no statistically significant relationship between the presence of these antibodies and either PMN recovery or incidence of transfusion reaction. We conclude that the prospective use of these assays is of little value in predicting the recipient's response to PMN transfusion.
GRANULOCYTE (PMN)transfusion therapy is of recognized clinical benefit in infected neutropenic patient^,'*^*^ and may be lifesaving in persistently granulocytopenic indiv i d u a l ~It. ~has been suggested that survival of PMNs in the transfused host is compromised by the presence of preformed alloantibodies directed against donor cells, as evidenced by decreased circulation of PMNs and by nonhemolytic transfusion reactions.e-8*'oAssays for the detection of antileukocyte antibodies in the form of leukoagglutinins, lymphocytotoxins and granulocytotoxins have been devised with the intent of providing information useful in selecting donors and avoiding transfusion reactions.* We have employed four such assays in an attempt to clarify the relationship between recipient presensitization and response to PMN transfusion. Supported in part by NIH contract RNOI-CM63810. Received for publication December 23. 1977; accepted February 20, 1978.
Materials and Methods One hundred and eighty-seven PMN transfusions given to 19 neutropenic cancer patients during 20 episodes of septicemia were evaluated. All patients had been entered in a randomized study of the utility of PMN transfusion therapy, and had been multiply transfused prior to the septic episode. PMNs were collected using the continuous flow centrifuge (CFC) in all instances. All transfusions were ABO compatible but were not HL-A matched. The median number of transfusions per recipient was six. Recipients were pretreated with diphenhydramine, hydrocortisone, acetominophen, and or chlorpromazine in 15 instances. Recipient sera were obtained twice weekly during the period of transfusion, stored at -70 C, and examined daily for the presence of antileukocyte antibody against the donor. The techniques for these assays, published previously,2 are briefly outlined below. Granulocytotoxicity ( G )
Serial twofold dilutions of heat-inactivated recipient serum were mixed with a standardized donor PMN suspension. Following a 3lXminute incubation at 23 C, complement was added and incubation continued for another 60 minutes. Eosin Y and formaldehyde were then added, the cells were allowed to settle, and the degree of granulocytotoxicity was ascertained employing an inverted phase contrast microscope. Controls .included a negative serum (the serum autologous to the PMNs), a known positive serum, and a buffer control. Lymphocytotoxicity ( L )
Serial twofold dilutions of heat-inactivated recipient serum were mixed with a standardized donor lymphocyte suspension collected at the lymphocyte-rich interface layer of FicollHypaque gradients. Following incubation at 23 C for 60 minutes, complement was added and lymphocytotoxicity determined in a manner identical to that outlined for granulocytotoxicity.
0041-1132/79/0100/0090$00.75 0 J. B. Lippincott CO. Transfusion January-February 1979
90
Volume 19 Number I
Volume 19 Number I
.
91
ANTILEUKOCYTE ANTIBODY SCREENING
Controls consisted of the serum autologous to the lymphocytes, a buffer control, and a known positive serum.
Table 1. Incidence of Positive Antileukocyte Antibody Assays
Assay
No. of Positive Samples
Per cent of Total Samples (n = 187)
Granulocytotoxicity Lymphocytotoxicity Microleukoagglutination Capillary leukoagglutination
5 6 9 87
3 3 5 47
MicroleukoaK~Iutination( M )
Serial twofold dilutions of heat-inactivated serum were mixed with a standardized donor leukocyte suspension and incubated at 37 C for two hours. Toluidene blue was then added and the mixtures were examined using an inverted light microscope. The degree of leukocyte agglutination was scored on a 0 (no agglutination) to 4+ (strong agglutination) scale. Controls were run as above. Capillary Leukoagglutination (C)
Serial twofold dilutions of heat-inactivated recipient serum were mixed with a standardized donor leukocyte suspension. Following I5 minutes of incubation at room temperature, capillary tubes were filled with aliquots of the reaction mixture, sealed at one end, and centrifuged. The capillary tubes were then maintained at a 45 degree angle with the leukocyte pellet uppermost, for 60 minutes at 23 C. The length of leukocyte migration (mm) was measured employing a phase contrast microscope. Each assay was run in tandem with both positive and negative control sera. Assays were set up in triplicate and the mean migration length of the leukocytes was employed for calculation of the per cent inhibition of leukocyte migration in comparison to migration in autologous (negative) serum. One hour posttransfusion PMN increments per l o l l transfused PMNs per square meter of body surface area (corrected 1 hour PMN increments) were calculated and compared with the assay results. The response to transfusion was not known to the individual performing the in vitro assays. Assays were considered positive if, at any dilution up to 1:512, there was greater than 10 per cent cell death or 1 + agglutination; or if there was 10 per cent inhibition of migration up to a dilution of 1: 16. Records of each PMN transfusion administered to these patients were examined for correlation between presence of antibody and occurrence of nonhemolytic transfusion reactions (chills, fever, dyspnea, cyanosis, and/or hypotension). At the time of transfusion, a nurse had been in attendance monitoring and recording vital signs every ten minutes. A physician was called to evaluate the patient in the event of suspected reaction, and the clinical impression recorded on a protocol sheet.
Results Ninety-six of the 187 transfusions (52%) were associated with the presence of antibodies against donor leukocytes as determined by one or more of the assays used. The majority of these antibodies was detected by the C assay (Table 1). Ten donor-recipient pairs (5%) had antibody demonstrated by more than one assay. There were six instances (3%) in which the C assay was negative while antibody was detected by one of the other methods. The mean corrected 1 hour granulocyte increment for all patients was 2149 (SD = 9073), with a range of -43,683 to +54,954. The extreme high and low increments were seen following transfusion in which small doses of PMNs were given to large patients and the increments then corrected to 10" transfused cells per square meter body surface area. Transfusions in which the recipient serum demonstrated no preformed antibody produced a mean corrected 1 hour PMN increment of 2803 (SD = 6637) while presence of antibody was associated with a mean increment of 1487 (SD = 11,OOO). Although the increment appeared to be greater in the group without preformed antibody, the within-group variance was extremely large (Fig. l), and the difference between the two groups was not statistically significant (p > 0.10; Wilcoxon Rank Sum Test). One hundred twenty-five transfusions resulted in PMN increments at 1 hour. The influence of preformed antileukocyte antibody on PMN recovery is presented in Table 2. Chi-square analysis revealed no statistically significant relationship between the presence of antibody and PMN recovery ( x 2 = 0.007, p > 0.10). Particular attention was given to the capillary leukoagglutination assay, as this technique most frequently detected antibody. In the presence of circulating leukoagglutinins as detected by this assay, 64 per cent of recipients achieved granulocyte increments compared to 66 per cent of leukoagglutinin-free recipients. This difference
92
UNGERLEIDER ET AL.
56,000
r
Transfusion January-February 1979
0
I-
FIG.1. Distribution of corrected granulocyte increments as a function of antileukocyte antibody assay results.
0
a 3
9 z"
0
-8,OOo
n
za -16,000 a
8 -24,000
-32,0004,OOo
-
1
I
was not statistically significant ( x ' p > 0.10).
I
= 0.0037,
Nonhemolytic transfusion reactions occurred in 9/187 (5%) of these transfusions. While none of the reactions was considered lifethreatening, three of them were characterized by respiratory compromise reflected in cyanosis of
I
I
the nailbeds and/or dyspnea, and were considered severe. As shown in Table 3, the presence of preformed circulating antileukocyte antibody detected by any of our assays did not influence the incidence of reaction ( x 2 = 0.75, p > 0.10). Separate analysis of the C assay (x' = 0.30, p > 0.10) and the G, L and M assays
Volume 19 Number 1
93
ANTILEUKOCYTE ANTIBODY SCREENING
Table 2. Influence of Preformed Antileukocyte Antibody on 1 Hour Granulocyte Recovery.
Antibody
No. Transfusions Producing Increments
No. Transfusions Producing No Change or Decrements
Present Absent
61 (33%) 64 (33%)
29 (16%) 33 (1 So/,)
( x 2 = 1.37,
p > 0.10) uncovered no statistically significant correlation. On two occasions, a recipient with a positive C assay was premedicated with hydrocortisone and diphenhydramine, and may have had a potential reaction aborted by these drugs. Adding these two patients to the group having a positive C assay with reaction, and subtracting them from the group having a positive C assay without reaction, did not reveal a statistically significant correlation ( x 2 = 1.37, p > 0.10). The distribution of nonhemolytic transfusion reactions as a function of levels of antibody as detected by the C assay is presented in Table 4. Chi-square analysis indicated that the relationship between per cent inhibition of migration of leukocytes and severity of reaction did not differ from chance (xz = 3.18, p > 0.10). This analysis was not extended to G, L, or M assays as only two reacting recipients had positive results .
Discussion
We have investigated the relationship between the presence of preformed antileukocyte antibody and the response of the recipient to transfused PMNs. Using a battery of four immunohematologic assays, we detected antibody against donor leukocytes in the majority of our recipients. We were unable, however, to demonstrate a significant relationship between the presence of such antibodies and either decreased recovery of circulating PMNs or increased incidence of nonhemolytic transfusion reactions. Thus, our findings indicate that the prospective use of these assays in screening for antileukocyte antibody is not of value in predicting successful donor-recipient pairs in terms of reaction-free transfusions or posttransfusion increments. It is admittedly difficult to evaluate the
Table 3. Influence of Preformed Antileukocyte Antibody on Incidence of Nonhemolytic Transfusion Reaction
No. of Transfusions (n = 187) Antibody Assay
Assay Result
With Reactions
Without Reactions
C, G, Lor M
Positive Negative
6 (3%) 3 (2%)
82 (44%) 96 (51%)
C
Positive Negative
5 (3%) 4 (2%)
72 (39%) 106 (56%)
G, L or M
Positive Negative
2 (1%) 7 (4%)
11 (6%) 167 (89%)
response to an individual granulocyte transfusion. We have used the occurrence of transfusion reactions and the 1 hour posttransfusion increment because they are clearly definable parameters and are clinically relevant. Reaction-free transfusions are obviously desirable, and the circulating granulocyte level is known to correlate with the incidence and severity of infectious episode^.^ Prior studies have suggested that presensitization is of consequence in patients receiving PMN replacement therapy. Wellcontrolled animal studies employing these assays have shown that prior alloimmunization results in decreased circulation of granulocytes.2 In addition, transfusion of leukoagglutinins has been associated with severe pulmonary reaction in humans.'O Although the results of the present study do not support these findings, there are mulTable 4. Effect of Per Cent Inhibition of Leukocyte Migration as Measured by C Assay on Incidence of Nonhemolytic Transfusion Reaction' Per cent Inhibition of Leukocyte Migration
0-9 10-30 40-60 70- 100
No. of Transfusions Associated with
No Reaction
Mild Reaction
Severe Reaction
99
3
47
2 1 0
2 1
19
13
0 0
UNGERLEIDER ET AL.
tiple factors which may obscure the interpretation of response to granulocyte transfusion in the usual clinical setting. Recipients are found to be heterogeneous, with different degrees of neutropenia and hypersplenism, which may affect circulation of granulocytes. Concurrent drug therapy may alter the patient’s ability to mount an immune response, or, as with steroids, change the rate of entry and egress of granulocytes from the b l o ~ d .In~ addition, there are technical limitations imposed when counting low levels of leukocytes, which are then compounded by similar limitations in doing differentials on leukopenic individuals. Small errors of measurement under these circumstances become magnified when calculations are performed to account for differences in body surface area. It becomes apparent, then, that the wide variations in recorded increments could mask the consequences of alloimmunization when increments are used as the criterion. Similarly, recognition of nonhemolytic transfusion reactions can be affected by differences in drug therapy, mental status, and previous blood component therapy, as well as awareness of the observer. It is possible that with larger doses of infused granulocytes, more sensitive techniques of measuring granulocyte circulation, and improved methods of detecting antibody, pretransfusion screening will be of importance. At the present time, however, given the limitations outlined above, the prospective use of these assays is of little clinical utility. References 1. Alavi, J. B., R. K. Root, I. Djerassi. A. E.
Evans, S. J. Gluckman, R. R. MacGregor, D. Guerry, A. D. Schreiber, J. M. Shaw, P. Koch, and R. A. Cooper: A randomized trial of granulocyte transfusions for infection in acute leukemia. N. Engl. J. Med. 296:706, 1977. 2. Appelbaum, F. R., R. J. Trapani, and R. G. Graw,
Trmsfhsion JMWFebruary 1979
Jr.: Consequences of prior alloimmunization during granulocyte transfusion. Transfusion 17: 460, 1977. 3. Bishop, C. R., J. W. Athens, D. R. Boggs, H. R.
Warner, G. E. Cartwright, and M. M.Wintrobe: Leukokinetic studies. XIII. A non-steady-state kinetic evaluation of the mechanism of cortisone-induced granulocytosis. J. Clin. Invest. 47: 249, 1968. 4. Bodey, G. P.,
5. 6.
7.
8.
M. Buckley, Y. S. Sathe, and E. J. Freireich: Quantitative relationships between circulating leukocytes and infection in patients with acute leukemia. Ann. Intern. Med. 64:328, 1966. Boggs, D. R.: Neutrophils in the blood bank. N. Engl. J. Med. 2%:748, 1977. Cooper, M. R., E. Heise, F. Richards, J. Kaufmann. and C. L. Spurn: A prospective study of histocompatible leukocyte and platelet transfusions during chemotherapeutic induction of acute myeloblastic leukemia. I n : Leukocytes: Separation, Collection and Transfusion. J. M. Goldman and A. M. Lowenthal, Eds. London, Academic Press, 1975, p. 436. Goldstein, I. M., H. J. Eyre, P. I. Terasaki, E. S. Henderson, and R. G. Graw, Jr.: Leukocyte transfusions: Role of leukocyte alloantibodies in determining transfusion response. Transfusion 11:19, 1971. Graw, R. G. Jr., G. Herzig, S. Perry, and E. S. Henderson: Normal granulocyte transfusion therapy. Treatment of septicemia due to gramnegative bacteria. N. Engl. J. Med. 282367,
1972. 9. Herzig, R. H., G. P. Herzig, R. G. Graw, Jr..
M. I. Bull, and K. K. Ray: Successful granulocyte transfusion therapy for gram-negative septicemia. N. End. J. Med. 2%:701, 1977. 10. Thompson, J. S., C. D. Severson, M.J. Parmely, B. L. Marmorstein, and A. Simmons: Pulmonary “hypersensitivity” reactions induced by transfusion of non-HL-A leukoagglutinins. N. Engl. J. Med. 284:1120, 1971.
Richard S. Ungerleider, M.D.,Investigator, Pediatric Oncology Branch, National Cancer Institute, Bethesda, MD 20014. Frederick R. Appelbaum, M.D., Investigator, Pediatric Oncology Branch. Robert-John Trapani, Director, Department of Immunology, Microbiological Associates, Bethesda, MD 20014.
Albert B. Deisseroth, M.D., Ph.D., Head, Experimental Hematology Section, Pediatric Oncology Branch.
