T R A N SO FR UI SG II ON NA LC OAMRPTLIICCLAET I O N S Transfusion and component characteristics are not associated with allergic transfusion reactions to apheresis platelets William J. Savage,1 Aaron A.R. Tobian,2 Jessica H. Savage,1 Robert G. Hamilton,3 P. Dayand Borge,4 Richard M. Kaufman,1 and Paul M. Ness2
BACKGROUND: Transfusion-related characteristics have been hypothesized to cause allergic transfusion reactions (ATRs) but they have not been thoroughly studied. The primary objective of this study is to evaluate the associations of infusion rate, infusion volume, ABO mismatching, component age, and pretransfusion medication with the incidence and severity of ATRs. A secondary objective is to compare the risk of these attributes relative to the previously reported risk factor for aeroallergen sensitization in transfusion recipients, as measured by an aeroallergen-specific immunoglobulin (Ig)E antibody screen. STUDY DESIGN AND METHODS: Clinical and transfusion-related data were collected on subjects with reported ATRs and uneventful (control) apheresis platelet (PLT) transfusions over a combined 21-month period at two academic medical centers. Control transfusions were selected as the next uneventful transfusion after an ATR was reported. Logistic regression, MannWhitney, and t tests were used to assess associations with ATRs. Previously reported aeroallergen-specific IgE screening data were incorporated into a multivariable logistic regression. RESULTS: A total of 143 ATRs and 61 control transfusions were evaluated among 168 subjects, ages 2 to 86 years. Infusion rate, infusion volume, ABO mismatching, component age, and pretransfusion medication showed no significant association with ATRs (p > 0.05). Neither infusion rate nor infusion volume increased the risk of anaphylaxis versus mucocutaneous-only ATRs. Aeroallergen sensitization has previously been associated with ATRs. After transfusion-related covariates were controlled for, aeroallergen sensitization remained significantly associated with ATRs (odds ratio, 2.68; 95% confidence interval, 1.26-5.69). CONCLUSIONS: Transfusion- and component-specific attributes are not associated with ATRs. An allergic predisposition in transfusion recipients is associated most strongly with ATR risk.
296 TRANSFUSION Volume 55, February 2015
A
llergic transfusion reactions (ATRs) are a common complication of blood transfusion, particularly platelet (PLT) transfusion. Overall, the incidence of ATRs to PLT transfusions is 1% to 4%.1,2 The plasma fraction of PLT components appears to harbor the allergic stimulus for ATRs.1,3,4 In unusual circumstances, recipient antibodies5-7 or passive transfusion of donor antibodies8-12 can cause ATRs. However, these risk factors do not explain most ATRs. A pervasive risk factor appears to be individuals with an allergic predisposition,13 specifically aeroallergen sensitization.14-17 Component- and transfusion-related factors may also influence the risk of ATRs, but they have not been thoroughly studied. Pretransfusion medication,18 transfusion rate, ABO compatibility,19 component age,20 and PLT source21 have been hypothesized to affect the risk of ATRs. These are modifiable risk factors, so understanding their role in ATRs is important for designing transfusion strategies that minimize the impact of ATRs.
ABBREVIATION: ATR(s) = allergic transfusion reaction(s). From the 1Department of Pathology, Brigham and Women’s Hospital, Boston, Massachusetts; the 2Department of Pathology and the 3Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland; and the 4Greater Chesapeake and Potomac American Red Cross, Baltimore, Maryland. Address reprint requests to: William J. Savage, Amory 260, 75 Francis Street, Boston, MA 02115; e-mail: [email protected]. Support by the National Institutes of Health (R21HL107828 and 1K23AI093152-01A1), the American Society of Hematology Scholar Award, and the Doris Duke Charitable Foundation Clinician Scientist Development Award (#22006.02). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Received for publication May 23, 2014; revision received July 23, 2014, and accepted July 23, 2014. doi: 10.1111/trf.12852 © 2014 AABB TRANSFUSION **;**:**-**. 2015;55:296–300. Volume **, ** **
TRANSFUSION
1
SAVAGE ET AL.
A
ALLERGIC TRANSFUSION REACTIONS
p = 0.5
B
p = 0.3
Statistical analysis
0.5
0.5
2
Infusion rate (mL/kg/min)
Infusion rate (mL/kg/min)
Summary statistics are presented as mean ± SD or median (interquartile range) for continuous variables and as 0.4 0.4 proportions for binary variables. Data 0.3 were log transformed for statistical 0.3 testing when transformation achieved 0.2 normality, as assessed by the Shapiro0.2 Wilk test. Unpaired, two-tailed t tests 0.1 with unequal variance were used to 0.1 compare continuous variables between 0 groups for normally distributed values. 0 Mann-Whitney U tests were used to Control ATR Mucocutaneous Anaphylaxis compare continuous values for Fig. 1. Infusion rates of apheresis PLTs and their association with incidence and nonnormally distributed variables. severity of ATRs. (A) Infusion rates (mL/kg/min) in uneventful transfusions (n = 51) Comparisons of proportions between and ATRs (n = 124). (B) Infusion rates in ATRs that were mucocutaneous-only groups were performed with a two(n = 98) versus ATRs that involved two or more systems (anaphylaxis, n = 26). tailed Fisher’s exact test. For logistic regression, a complete case analysis was used (n = 158 observations: 45 controls MATERIALS AND METHODS [74%] included; 113 ATR cases [79%] included) with robust Study design variance estimation. Significance was defined as p values The study design, population, and methods have been of less than 0.05 and a confidence interval (CI) that does described previously.16 Briefly, patients with ATRs not include 1, for odds ratio (OR) estimates. Analysis was reported to the blood bank at Johns Hopkins Hospital conducted with computer software (Stata, Version 13.1, from July 2011 to May 2012 and Brigham and Women’s StataCorp, College Station, TX). Hospital/Dana Farber Cancer Institute from November 2012 to August 2013 were invited to enroll. Previously RESULTS enrolled subjects who experienced subsequent ATRs during the study period were included as additional ATR PLT transfusions cases. The study was institutional review board approved During the study period, 143 ATRs and 61 control transfuat both institutions, and informed consent was obtained sions were evaluated among 168 subjects, ages 2 to 86 from all subjects. Control transfusions were selected as the years. Criteria for anaphylaxis were met in 26 ATRs patient receiving the next PLT component issued to a non(18.2%). Among all subjects, 59% were male (n = 99). Six surgical location after an ATR was observed. The enrollPLT components were concentrated, four of which caused ment ratio of ATR:control subjects was targeted at 2:1. The ATRs. Two components were washed, both of which study was designed to have 80% power to detect a 20% caused ATRs. In 41 transfusions (28 ATRs and 13 control), absolute difference in atopic disease prevalence between two PLT components were transfused consecutively as a ATR subjects and controls. Medical histories were continuous infusion. In these instances, the second comobtained by a research coordinator via structured patient ponent was a split from the same donation in 24 transfu(and family, as appropriate) questionnaires and chart sions. Of subjects with at least one ATR and two PLT reviews. Allergic histories and ATR manifestations have transfusions before enrollment, 33% (34/103) had at least been reported previously.16 Infusion rate, total volume one additional ATR. transfused, ABO typing, PLT age, and pretransfusion medication data were collected at the time of enrollment. Transfusion rate and volume Reaction severity was coded as mucocutaneous (i.e., skin and/or mucosal symptoms only) or anaphylaxis (i.e., The mean infusion rate of apheresis PLTs was mucocutaneous plus respiratory, cardiovascular, or gas6.3 ± 3.5 mL/min. Mean weight-normalized infusion rates trointestinal symptoms), according to consensus criteria22 were 0.094 ± 0.074 mL/kg/min. Mean infusion rates were 40% higher at Brigham and Women’s Hospital/Dana and review by an allergist. All PLT components were colFarber Cancer Institute than Johns Hopkins Hospital lected by apheresis. Plasma aeroallergen–specific immu(0.11 mL/kg/min vs. 0.081 mL/kg/min, respectively; noglobulin (Ig)E multiallergen screen analysis (Phadiatop, p = 0.005). Figure 1A shows the infusion rates (mL/kg/ ThermoFisher Scientific/Phadia, Kalamazoo, MI) has min) in apheresis PLT transfusions that did or did not been reported previously.16 A clinically positive test was result in an ATR. Mean infusion rates were 8% slower in defined as at least 0.35 kUa/L. TRANSFUSION
Volume **, ** **
Volume 55, February 2015 TRANSFUSION 297
ALLERGIC TRANSFUSION REACTIONS
SAVAGE ET AL.
the ATR group than in the control group (0.092 mL/kg/ min vs. 0.10 mL/kg/min, respectively; p = 0.5). To evaluate whether infusion rate is related to severity of ATRs, we compared ATR infusion rates in mucocutaneous-only reactions to anaphylaxis (Fig. 1B). Mean infusion rates were 29% higher in anaphylactic versus mucocutaneousonly reactions, but the difference was not significant (0.087 mL/kg/min vs. 0.11 mL/kg/min; p = 0.3). In addition to infusion rate, total infusion volume could plausibly be a risk factor for ATRs if the allergic stimulus is dose dependent. Subjects who received two consecutive PLT products represent a group who receive twice as much component volume as the rest of the cohort. However, a similar proportion of controls received twice as many transfusions (13 of 61, 21.3%) as those who experienced ATRs (28 of 143, 19.6%, p = 0.8). We also evaluated the total transfusion volume on a mL/kg basis for all subjects. The total volume transfused was similar between the ATR and control groups (mean, 4.5 mL/kg vs. 4.2 mL/kg, respectively; p = 0.2). The volume transfused was not different between mucocutaneous-only and anaphylactic ATRs (p = 0.8).
ABO-mismatched transfusion ABO-mismatched transfusion was assessed as major (cellular incompatibility), minor (plasma incompatibility), or any mismatch. Of all transfusions, 48% were ABO mismatched. Table 1 shows the proportion of ATR and control transfusions that were mismatched. There were no significant differences between ATR and control groups (Table 1).
before transfusion (Table 2). In some cases, patients received these drugs for underlying medical conditions, and medication administration coincidentally preceded transfusion. Thus, diphenhydramine dosing within 6 hours before transfusion and H2 receptor antagonist or glucocorticoid dosing within 24 hours before transfusion was considered a pretransfusion medication. There were no significant differences in pretransfusion medication frequency, although patients with ATRs tended to receive diphenhydramine pretransfusion medication more often. Two percent of subjects (4 of 204) received an H1 receptor antagonist other than diphenhydramine in the 24 hours preceding transfusion (two control—loratadine, hydroxyzine; two ATR—both loratadine).
Logistic model of ATR risk In univariable logistic regression, the following covariates were analyzed: infusion rate (mL/min), any ABO mismatch, component age, diphenhydramine pretransfusion medication, and glucocorticoid pretransfusion medication. None of these variables was significantly associated with ATRs (p > 0.05 for all analyses, Table 3). Similar results were seen in multivariable analysis. Addition of an interaction term to measure the effect of combined diphenhydramine and glucocorticoid pretransfusion medication was not significant in multivariable analysis, which included all covariates (p = 0.6). A previous report from this cohort identified that concentrations of aeroallergen-specific IgE (Phadiatop) are
TABLE 2. Documented pretransfusion medication*
Storage duration The median age of PLTs in the study was 5 days. PLT storage age was not different between transfusions that did or did not result in an ATR (p = 1). There was no difference in the proportion of younger PLTs between the ATR and control groups. PLT components causing ATRs were 2 or 3 days old in 14.3% of units, compared to 9.8% of control units (p = 0.5).
We assessed the frequency of diphenhydramine, H2 receptor antagonist, and glucocorticoid administration TABLE 1. ABO mismatch in ATR and control transfusions* No ATR (n = 61) 11 (18) 25 (41) 32 (52.5)
Control 17 (27.9) 10 (16.4) 11 (18)
ATR 60 (42.3) 33 (23.4) 16 (11.3)
p value 0.06 0.3 0.3
* Data are reported as number (%).
BACKGROUND: Transfusion-related characteristics have been hypothesized to cause allergic transfusion reactions (ATRs) but they have not been thoroughly studied. The primary objective of this study is to evaluate the associations of infusion rate, infusion volume, ABO mismatching, component age, and pretransfusion medication with the incidence and severity of ATRs. A secondary objective is to compare the risk of these attributes relative to the previously reported risk factor for aeroallergen sensitization in transfusion recipients, as measured by an aeroallergen-specific immunoglobulin (Ig)E antibody screen. STUDY DESIGN AND METHODS: Clinical and transfusion-related data were collected on subjects with reported ATRs and uneventful (control) apheresis platelet (PLT) transfusions over a combined 21-month period at two academic medical centers. Control transfusions were selected as the next uneventful transfusion after an ATR was reported. Logistic regression, MannWhitney, and t tests were used to assess associations with ATRs. Previously reported aeroallergen-specific IgE screening data were incorporated into a multivariable logistic regression. RESULTS: A total of 143 ATRs and 61 control transfusions were evaluated among 168 subjects, ages 2 to 86 years. Infusion rate, infusion volume, ABO mismatching, component age, and pretransfusion medication showed no significant association with ATRs (p > 0.05). Neither infusion rate nor infusion volume increased the risk of anaphylaxis versus mucocutaneous-only ATRs. Aeroallergen sensitization has previously been associated with ATRs. After transfusion-related covariates were controlled for, aeroallergen sensitization remained significantly associated with ATRs (odds ratio, 2.68; 95% confidence interval, 1.26-5.69). CONCLUSIONS: Transfusion- and component-specific attributes are not associated with ATRs. An allergic predisposition in transfusion recipients is associated most strongly with ATR risk.
296 TRANSFUSION Volume 55, February 2015
A
llergic transfusion reactions (ATRs) are a common complication of blood transfusion, particularly platelet (PLT) transfusion. Overall, the incidence of ATRs to PLT transfusions is 1% to 4%.1,2 The plasma fraction of PLT components appears to harbor the allergic stimulus for ATRs.1,3,4 In unusual circumstances, recipient antibodies5-7 or passive transfusion of donor antibodies8-12 can cause ATRs. However, these risk factors do not explain most ATRs. A pervasive risk factor appears to be individuals with an allergic predisposition,13 specifically aeroallergen sensitization.14-17 Component- and transfusion-related factors may also influence the risk of ATRs, but they have not been thoroughly studied. Pretransfusion medication,18 transfusion rate, ABO compatibility,19 component age,20 and PLT source21 have been hypothesized to affect the risk of ATRs. These are modifiable risk factors, so understanding their role in ATRs is important for designing transfusion strategies that minimize the impact of ATRs.
ABBREVIATION: ATR(s) = allergic transfusion reaction(s). From the 1Department of Pathology, Brigham and Women’s Hospital, Boston, Massachusetts; the 2Department of Pathology and the 3Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland; and the 4Greater Chesapeake and Potomac American Red Cross, Baltimore, Maryland. Address reprint requests to: William J. Savage, Amory 260, 75 Francis Street, Boston, MA 02115; e-mail: [email protected]. Support by the National Institutes of Health (R21HL107828 and 1K23AI093152-01A1), the American Society of Hematology Scholar Award, and the Doris Duke Charitable Foundation Clinician Scientist Development Award (#22006.02). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Received for publication May 23, 2014; revision received July 23, 2014, and accepted July 23, 2014. doi: 10.1111/trf.12852 © 2014 AABB TRANSFUSION **;**:**-**. 2015;55:296–300. Volume **, ** **
TRANSFUSION
1
SAVAGE ET AL.
A
ALLERGIC TRANSFUSION REACTIONS
p = 0.5
B
p = 0.3
Statistical analysis
0.5
0.5
2
Infusion rate (mL/kg/min)
Infusion rate (mL/kg/min)
Summary statistics are presented as mean ± SD or median (interquartile range) for continuous variables and as 0.4 0.4 proportions for binary variables. Data 0.3 were log transformed for statistical 0.3 testing when transformation achieved 0.2 normality, as assessed by the Shapiro0.2 Wilk test. Unpaired, two-tailed t tests 0.1 with unequal variance were used to 0.1 compare continuous variables between 0 groups for normally distributed values. 0 Mann-Whitney U tests were used to Control ATR Mucocutaneous Anaphylaxis compare continuous values for Fig. 1. Infusion rates of apheresis PLTs and their association with incidence and nonnormally distributed variables. severity of ATRs. (A) Infusion rates (mL/kg/min) in uneventful transfusions (n = 51) Comparisons of proportions between and ATRs (n = 124). (B) Infusion rates in ATRs that were mucocutaneous-only groups were performed with a two(n = 98) versus ATRs that involved two or more systems (anaphylaxis, n = 26). tailed Fisher’s exact test. For logistic regression, a complete case analysis was used (n = 158 observations: 45 controls MATERIALS AND METHODS [74%] included; 113 ATR cases [79%] included) with robust Study design variance estimation. Significance was defined as p values The study design, population, and methods have been of less than 0.05 and a confidence interval (CI) that does described previously.16 Briefly, patients with ATRs not include 1, for odds ratio (OR) estimates. Analysis was reported to the blood bank at Johns Hopkins Hospital conducted with computer software (Stata, Version 13.1, from July 2011 to May 2012 and Brigham and Women’s StataCorp, College Station, TX). Hospital/Dana Farber Cancer Institute from November 2012 to August 2013 were invited to enroll. Previously RESULTS enrolled subjects who experienced subsequent ATRs during the study period were included as additional ATR PLT transfusions cases. The study was institutional review board approved During the study period, 143 ATRs and 61 control transfuat both institutions, and informed consent was obtained sions were evaluated among 168 subjects, ages 2 to 86 from all subjects. Control transfusions were selected as the years. Criteria for anaphylaxis were met in 26 ATRs patient receiving the next PLT component issued to a non(18.2%). Among all subjects, 59% were male (n = 99). Six surgical location after an ATR was observed. The enrollPLT components were concentrated, four of which caused ment ratio of ATR:control subjects was targeted at 2:1. The ATRs. Two components were washed, both of which study was designed to have 80% power to detect a 20% caused ATRs. In 41 transfusions (28 ATRs and 13 control), absolute difference in atopic disease prevalence between two PLT components were transfused consecutively as a ATR subjects and controls. Medical histories were continuous infusion. In these instances, the second comobtained by a research coordinator via structured patient ponent was a split from the same donation in 24 transfu(and family, as appropriate) questionnaires and chart sions. Of subjects with at least one ATR and two PLT reviews. Allergic histories and ATR manifestations have transfusions before enrollment, 33% (34/103) had at least been reported previously.16 Infusion rate, total volume one additional ATR. transfused, ABO typing, PLT age, and pretransfusion medication data were collected at the time of enrollment. Transfusion rate and volume Reaction severity was coded as mucocutaneous (i.e., skin and/or mucosal symptoms only) or anaphylaxis (i.e., The mean infusion rate of apheresis PLTs was mucocutaneous plus respiratory, cardiovascular, or gas6.3 ± 3.5 mL/min. Mean weight-normalized infusion rates trointestinal symptoms), according to consensus criteria22 were 0.094 ± 0.074 mL/kg/min. Mean infusion rates were 40% higher at Brigham and Women’s Hospital/Dana and review by an allergist. All PLT components were colFarber Cancer Institute than Johns Hopkins Hospital lected by apheresis. Plasma aeroallergen–specific immu(0.11 mL/kg/min vs. 0.081 mL/kg/min, respectively; noglobulin (Ig)E multiallergen screen analysis (Phadiatop, p = 0.005). Figure 1A shows the infusion rates (mL/kg/ ThermoFisher Scientific/Phadia, Kalamazoo, MI) has min) in apheresis PLT transfusions that did or did not been reported previously.16 A clinically positive test was result in an ATR. Mean infusion rates were 8% slower in defined as at least 0.35 kUa/L. TRANSFUSION
Volume **, ** **
Volume 55, February 2015 TRANSFUSION 297
ALLERGIC TRANSFUSION REACTIONS
SAVAGE ET AL.
the ATR group than in the control group (0.092 mL/kg/ min vs. 0.10 mL/kg/min, respectively; p = 0.5). To evaluate whether infusion rate is related to severity of ATRs, we compared ATR infusion rates in mucocutaneous-only reactions to anaphylaxis (Fig. 1B). Mean infusion rates were 29% higher in anaphylactic versus mucocutaneousonly reactions, but the difference was not significant (0.087 mL/kg/min vs. 0.11 mL/kg/min; p = 0.3). In addition to infusion rate, total infusion volume could plausibly be a risk factor for ATRs if the allergic stimulus is dose dependent. Subjects who received two consecutive PLT products represent a group who receive twice as much component volume as the rest of the cohort. However, a similar proportion of controls received twice as many transfusions (13 of 61, 21.3%) as those who experienced ATRs (28 of 143, 19.6%, p = 0.8). We also evaluated the total transfusion volume on a mL/kg basis for all subjects. The total volume transfused was similar between the ATR and control groups (mean, 4.5 mL/kg vs. 4.2 mL/kg, respectively; p = 0.2). The volume transfused was not different between mucocutaneous-only and anaphylactic ATRs (p = 0.8).
ABO-mismatched transfusion ABO-mismatched transfusion was assessed as major (cellular incompatibility), minor (plasma incompatibility), or any mismatch. Of all transfusions, 48% were ABO mismatched. Table 1 shows the proportion of ATR and control transfusions that were mismatched. There were no significant differences between ATR and control groups (Table 1).
before transfusion (Table 2). In some cases, patients received these drugs for underlying medical conditions, and medication administration coincidentally preceded transfusion. Thus, diphenhydramine dosing within 6 hours before transfusion and H2 receptor antagonist or glucocorticoid dosing within 24 hours before transfusion was considered a pretransfusion medication. There were no significant differences in pretransfusion medication frequency, although patients with ATRs tended to receive diphenhydramine pretransfusion medication more often. Two percent of subjects (4 of 204) received an H1 receptor antagonist other than diphenhydramine in the 24 hours preceding transfusion (two control—loratadine, hydroxyzine; two ATR—both loratadine).
Logistic model of ATR risk In univariable logistic regression, the following covariates were analyzed: infusion rate (mL/min), any ABO mismatch, component age, diphenhydramine pretransfusion medication, and glucocorticoid pretransfusion medication. None of these variables was significantly associated with ATRs (p > 0.05 for all analyses, Table 3). Similar results were seen in multivariable analysis. Addition of an interaction term to measure the effect of combined diphenhydramine and glucocorticoid pretransfusion medication was not significant in multivariable analysis, which included all covariates (p = 0.6). A previous report from this cohort identified that concentrations of aeroallergen-specific IgE (Phadiatop) are
TABLE 2. Documented pretransfusion medication*
Storage duration The median age of PLTs in the study was 5 days. PLT storage age was not different between transfusions that did or did not result in an ATR (p = 1). There was no difference in the proportion of younger PLTs between the ATR and control groups. PLT components causing ATRs were 2 or 3 days old in 14.3% of units, compared to 9.8% of control units (p = 0.5).
We assessed the frequency of diphenhydramine, H2 receptor antagonist, and glucocorticoid administration TABLE 1. ABO mismatch in ATR and control transfusions* No ATR (n = 61) 11 (18) 25 (41) 32 (52.5)
Control 17 (27.9) 10 (16.4) 11 (18)
ATR 60 (42.3) 33 (23.4) 16 (11.3)
p value 0.06 0.3 0.3
* Data are reported as number (%).
