Storage of washed or irradiated red cells in AS-7 improves their in vitro characteristics.

BACKGROUND: AS-7 is a new alkaline hypotonic red cell additive solution (AS) shown to improve red cell quality during storage compared with AS-1. We sought to compare red cells stored in AS-7 with those stored in SAGM using RCC that were either untreated, or washed or irradiated on day 14 of storage. STUDY DESIGN AND METHODS: A pooled and split study design was used to produce seven identical RCC (four in SAGM and three in AS-7). At day 14 following donation, two RCC (one in SAGM and one in AS-7) were gamma irradiated and three RCC (two in SAGM and one in AS-7) were washed and resuspended in either SAGM or AS-7. RCC were sampled for analysis throughout storage and at end of shelf life: day 28 for washed or irradiated and day 35 for untreated RCC. RESULTS: For untreated, washed or irradiated RCC, those stored in AS-7 had lower haemolysis, red cell microvesicles and supernatant potassium content than RCC in SAGM. In addition, ATP levels and pH were better maintained in AS-7 RCC than in SAGM RCC. CONCLUSION: These data suggest that the quality of these components may be improved by storage in AS-7 compared with SAGM.

Autologous platelet-rich plasma isolated using the Haemonetics Cell Saver 5 and Haemonetics MCS+ for the preparation of platelet gel.

BACKGROUND AND OBJECTIVES: We compared three methods of isolating platelet-rich plasma (PRP) using the Haemonetics Cell Saver 5 and one method of isolating PRP by plateletpheresis using the Haemonetics MCS+. PRP contains both platelets and fibrinogen, which are used in the preparation of haemostatic agents. MATERIALS AND METHODS: When the Haemonetics Cell Saver 5 was used, 500 ml of blood from each of 30 normal volunteer donors was collected into 70 ml of citrate-phosphate-dextrose (CPD) anticoagulant. In a further 14 normal volunteers, the Haemonetics MCS+ was used to isolate PRP by plateletpheresis using an acid citrate dextrose (ACD) to blood ratio of 1 : 9. In a separate study, CPD-anticoagulated whole blood from another 30 volunteers was used for measurement of fibrinogen levels in the plasma and cryoprecipitate. RESULTS: A larger volume of PRP can be collected using the Haemonetics Cell Saver 5 than by using the Haemonetics MCS+. The platelet concentration and the total number of platelets were higher in the PRP isolated using the Haemonetics MCS+ than in the PRP isolated by the three methods used with the Haemonetics Cell Saver 5, with differences in platelet concentration and PRP volume among the four methods. The mean fibrinogen level in the plasma was 253 mg % +/- 47 (SD) and in the cryoprecipitate was 1085 mg % +/- 304 (SD). CONCLUSIONS: The most appropriate method of PRP isolation for preparation of platelet gel is dependent upon the specific surgical procedure to be undertaken and the patient's needs.

Transfusion and lung injury.

The respiratory tree has been viewed as an infrequent site of injury arising as a complication of transfusion. In recent years, this view has changed as investigators have shown that two complications--circulatory overload and transfusion-related acute lung injury--are relatively frequent events. Circulatory overload is a result of hypertransfusion to individuals at risk, the very young or old recipient. The reaction is due to fluid infusion which overwhelms the capacity of the left ventricle, resulting in pulmonary edema. While rarely fatal, studies have shown that such incidents result in intensive care and extended hospitalization. In the setting of orthopedic surgery, 1% of elderly patients undergoing hip or knee surgery experience circulatory overload. These events are associated with autologous, as well as allogeneic red blood cells (RBC) and fresh frozen plasma. Transfusionists need to be vigilant with transfusion therapy in this population. Phlebotomy and supplemental oxygen are the key therapies. Transfusion-related acute lung injury (TRALI) is the adult respiratory distress syndrome due to transfusion. It is associated with a significant morbidity and mortality of 5-14%, making it the third most common cause of death from transfusion in developed countries. It is characterized by the onset of acute respiratory distress, bilateral pulmonary edema and hypoxemia. It occurs within 1-2 hours of transfusion of a plasma-containing blood product. All blood components have been associated with the reaction, and rarely, intravenous immune globulin. There is no recognized profile of individuals at increased risk for TRALI. There are two purported mechanisms of injury; the vast majority of cases are associated with passively transfused complement-activating antibodies. These antibodies are either HLA (Class I or II) or granulocyte-specific. These antibodies appear to act as mediators, which result in granulocyte aggregation, activation, and microvascular pulmonary injury. With appropriate respiratory intervention, 80% of patients recover within 96 hours of the original insult. There are no permanent pulmonary sequelae.

Bigger is better: maternal and neonatal predictors of hematopoietic potential of umbilical cord blood units.

Umbilical cord blood (CB) is a useful stem cell source for patients without matched family donors. CB banking is expensive, however, because only a small percentage of the cord units stored are used for transplantation. In this study, we determined whether maternal factors, such as race, age, and smoking status have an effect on laboratory parameters of hematopoietic potential, such as viability, cell counts, CD34+ cell counts, and CFU-GM. We studied the effect of neonatal characteristics such as birth order, birth weight, gestational age, and sex of the baby on the same laboratory parameters. Race and maternal age had no effect on these laboratory parameters. In multivariate analysis, babies of longer gestational age had higher cell counts, but lower CD34+ cell counts and CFU-GM. Bigger babies had higher cell counts, more CD34+ cells, and more CFU-GM. Women with fewer previous live births also produced cord units with higher cell counts, CFU-GM, and CD34+ cell counts. Specifically, each 500 g increase in birth weight contributed to a 28% increase in CD34+ cell counts, each week of gestation contributed to a 9% decrease in CD34+cell counts, and each previous birth contributed to a 17% decrease in CD34+ cell counts (all P < 0.05). These data may be used to select the optimal cord blood donors and allow CB banks efficient resource allocation.

Transfusion to blood group A and O patients of group B RBCs that have been enzymatically converted to group O.

BACKGROUND: The transfusion of ABO-incompatible RBCs is the leading cause of fatal transfusion reactions. Group O RBCs, lacking terminal immunodominant A and B sugars to which humans are immunized, are safe for transfusion to persons of any ABO blood group. With the use of a recombinant alpha-galactosidase to remove terminal galactose from group B RBCs, the safety and efficacy of enzyme-converted group-B-to-group-O (ECO) RBC components were studied in transfusion-dependent patients. STUDY DESIGN AND METHODS: Twenty-four patients (blood groups A and O) were randomly assigned to receive transfusion(s) of either ECO or control group O RBCs. If a second transfusion was given, the other blood component was administered. RESULTS: Twenty-one patients were given ECO RBCs; 18 also underwent control transfusions. One patient received only a small aliquot for RBC survival studies, instead of a full-unit transfusion, because his serum was incompatible with ECO RBCs. No adverse events occurred. Both ECO and control transfusions resulted in appropriate Hb increments and comparable (51)Cr-labeled RBC survival studies. One patient developed a transient, weak-positive DAT, without hemolysis. Two weeks after transfusion, 5 of 19 evaluable ECO RBC recipients had increases in anti-B titers. CONCLUSION: ECO RBCs were comparable to group O cells for safety and efficacy in this study. The clinical significance of the increase in anti-B and of occasional serologic incompatibilities with ECO RBCs is unclear. If strategies can be developed to remove A epitopes, enzymatic conversion could be used to create a universal (group O) donor blood supply.

A case-controlled multicenter study of vasovagal reactions in blood donors: influence of sex, age, donation status, weight, blood pressure, and pulse.

BACKGROUND: Vasovagal reactions occur in a small, but significant number of blood donors. These reactions may decrease return donation and disrupt blood collection activities. The purpose of this study was to define the contributory role of sex, age, weight, blood pressure, and pulse in vasovagal reactions with syncope in blood donors. STUDY DESIGN AND METHODS: A retrospective case-control study involved 1890 blood donors with syncope from three large United States blood centers during 1994 and 1995. Case controls and random population controls were used in a logistic regression analysis to determine the significance of individual variables to syncopal reactions. RESULTS: Female donors, young donors, first-time donors, low-weight donors, and donors with low predonation blood pressure had higher absolute donation reaction rates than other donors. When each variable was adjusted for other variables by regression analysis, age, weight, and donation status (first-time or repeat donor) were significant (p<0.0001), and sex, predonation blood pressure, and predonation pulse were not. The most important variables, in descending order, were age, weight, and donation status (first-time or repeat donor). CONCLUSIONS: Donation-related vasovagal syncopal reactions are a multifactorial process determined largely by age, weight, and first-time donor status.

Predictive value of past and current screening tests for syphilis in blood donors: changing from a rapid plasma reagin test to an automated specific treponemal test for screening.

BACKGROUND: This study evaluated the change from a rapid plasma reagin (RPR) test to an automated specific treponemal test (PK-TP) in screening for syphilis in blood donors. STUDY DESIGN AND METHODS: A cross-sectional seroprevalence analysis was performed on 4,878,215 allogeneic blood donations from 19 American Red Cross Blood Services regions from May 1993 through September 1995. Positive predictive values relative to the confirmatory fluorescent treponemal antibody absorption test (FTA-ABS) were calculated. Differences in seroprevalence were compared in RPR and PK-TP tests for 1) unconfirmed and confirmed tests, 2) first-time and repeat donors, and 3) "recent" versus "past" infections. Donation data from three additional Red Cross regions were evaluated for repeat donation patterns of blood donors who had a donation that was positive in a serologic screening test for syphilis. The value of RPR and PK-TP tests as surrogate markers for HIV infection was compared. RESULTS: Reactive rates were lower but the positive predictive values was higher for the PK-TP test than for the RPR test. Initially, donors screened by PK-TP were more likely to be confirmed as positive than were donors screened by RPR, but these rates became comparable. It is estimated that a single HIV window-period donation was removed by serologic testing for syphilis each year of this study period. CONCLUSIONS: The change to the PK-TP test resulted in a lower repeatedly reactive rate, better prediction that a confirmed-positive test for syphilis would occur in testing in the FTA-ABS, fewer donations lost, and comparable deferral rates. Because of the high rate of reactivity to serologic testing for syphilis among donors previously confirmed positive for syphilis, indefinite deferral after a confirmed-positive index donation may be warranted. Serologic testing for syphilis is ineffective as a marker of HIV-infectious window-period donations.

Transfusion-associated graft-versus-host disease in an immunocompetent patient following cardiac surgery.

OBJECTIVES: We determined which of the 22 blood components obtained from unrelated donors and transfused to an apparently immunocompetent patient following open heart surgery caused transfusion-associated graft-versus-host disease (TA-GVHD). METHODS: Serologic and molecular methods were used to type the donors, the patient's family members, and the patient's postmortem tissues for HLA and a genetic marker on chromosome 17. RESULTS: Two donors were homozygous for the HLA class I antigens A1 B8, for which the patient was heterozygous. Both donors were heterozygous, not homozygous as expected, for the class II alleles. One of them had the same class II alleles as the patient (DRB1*0301, DRB3*0101/DRB1*0404, DRB4*0103). The patient's tissues were chimeric for restriction fragments at 17p13 of this donor. CONCLUSION: One-way HLA match leading to TA-GVHD can be caused by donor blood that is homozygous for class I and heterozygous for class II alleles. Two blood components given to our patient had such one-way HLA match. Class II alleles of the lymphocytes in one component were identical with those of the recipient and caused TA-GVHD. Class II alleles of the lymphocytes in the other component differed from those of the recipient and were eliminated either by the immune system of the patient or the lymphocytes that caused the TA-GVHD (graft versus graft).

Red blood cell transfusion practices in patients undergoing orthopedic surgery: a multi-institutional analysis.

This retrospective review analyzed and compared transfusion practices in patients undergoing orthopedic surgery in five Massachusetts hospitals with current practice guidelines; opportunities for improvement were identified. Patient-specific clinical information and data about transfusion practices were obtained from the medical records of 384 Medicare patients undergoing orthopedic surgery between January 1992 and December 1993. The number of patients who donated autologous blood preoperatively differed significantly among hospitals as did the number of autologous units that were unused. The number of blood units transfused at each transfusion event also differed significantly; some surgeons transfused > or =2 units in the majority of their patients, while others transfused 1 unit at a time. This variation in practice was not explained by differences in patients' clinical status. The mean pretransfusion hematocrit was higher for autologous versus allogeneic blood, suggesting more liberal criteria to transfuse autologous blood. Nearly half of all transfusion events were determined to have been potentially avoidable. Avoidable transfusions were also three to seven times more likely with autologous than with allogeneic blood. Significant inter-hospital differences existed in the number of elective surgery patients exposed to allogeneic blood. The major determinant of allogeneic blood use in these patients was the availability of autologous blood. Each additional autologous blood unit available decreased the odds of allogeneic blood exposure twofold. Differences in intraoperative and postoperative blood salvage use also were noted. These findings indicate that significant variations in practice exist. Comparative data enabled hospitals to identify and target specific areas for improvement.

Nationwide survey of home transfusion practices.

BACKGROUND: Limited information exists on home transfusion practices. STUDY DESIGN AND METHODS: In 1995, a survey requesting data for 1994 was sent to 1273 American Association of Blood Banks (AABB) institutional members and 113 non-AABB home health care agencies that provide out-of-hospital transfusions. RESULTS: Of 943 respondents, 102 provide blood to a home transfusion program, 37 provide blood and run a home transfusion program, and 13 run a home transfusion program only, for a total of 152 (16%) with some involvement in home blood transfusions. Most of the 50 respondents with a home transfusion program are licensed by their state and accredited by the Joint Commission on Accreditation of Healthcare Organizations. All respondents have written policies for home transfusion, and 90 percent require a signed informed-consent document before initiating transfusions in the home. Most have policies requiring that there be a second adult and a telephone in the home, that the home be deemed safe for transfusion, that the patient's physician be readily available, and that the patient have had prior transfusions. The most common component issued by the blood providers was red cells, followed by platelets. White cell-reduced components were always provided by 36 percent of respondents. The most common patient diagnosis was cancer. Home transfusions were provided primarily by registered nurses. Only 14 percent of respondents indicated that the medical director of the blood bank is responsible for approving a patient for home transfusion. A posttransfusion visit is performed by 46 percent of respondents. CONCLUSION: Although most facilities have policies for the administration of home transfusions, there remains marked heterogeneity among blood providers and transfusionists regarding home transfusion practices.

The association of biologically active lipids with the development of transfusion-related acute lung injury: a retrospective study.

BACKGROUND: Transfusion-related acute lung injury (TRALI) is clinically similar to the adult respiratory distress syndrome (ARDS) and has been linked to the transfusion of leukocyte antibodies in blood components. Animal model have implicated neutrophil (PMN)-priming agents in ARDS; however, two agents were required. Previous studies showed the generation of PMN-priming agents during blood storage. Thus the association of PMN-priming agents with TRALI was examined. STUDY DESIGN AND METHODS: Ten patients with TRALI and 10 with febrile or urticarial reactions (control group) were evaluated. The presence of PMN-priming activity was tested in the patients' pretransfusion and posttransfusion blood samples by incubating PMNs with these samples followed by activation of the respiratory burst. Plasma lipids were separated by normal-phase high-performance liquid chromatography (HPLC), and the priming activity was evaluated. The presence of leukocyte antibodies was determined in the blood donors and patients with TRALI. RESULTS: Significantly more PMN-priming activity was present in the posttransfusion sera (11.4 +/- 1.8 nmol superoxide anion/min, mean +/- SEM; n = 10) and plasma of patients with TRALI than in their pretransfusion sera (6.5 +/- 1.5: n = 10) or in the pretransfusion and posttransfusion sera (5.1 +/- 1.3, n = 10; and 4.5 +/- 1.4, n = 10, respectively) and from the controls (p < 0.05). HPLC separation of lipids demonstrated that three active species were present in the posttransfusion plasma samples of TRALI patients. All the patients with TRALI had underlying clinical factors, such as infection, cytokine administration, recent surgery, or massive transfusion, while only 2 of 10 control patients had these clinical conditions. None of the donors had significant titers of HLA or HLA-DR antibodies; however, 50 percent had weak positivity for granulocyte antibodies. CONCLUSION: TRALI is the result of two clinical events, the first being a predisposing clinical condition and the second being the transfusion of biologically active lipids in stored blood.

Limited utility of alanine aminotransferase screening of hepatitis C antibody-screened blood donors.

BACKGROUND: The introduction of hepatitis C virus (HCV) screening has significantly reduced the frequency of posttransfusion hepatitis C. To examine the current added value of alanine aminotransferase (ALT) screening, all donor screening at two large blood centers was reviewed. STUDY DESIGN AND METHODS: From July 1991 through March 1994, 1,258,000 allogeneic blood donors were screened by enzyme immunoassay for anti-HCV: 343,000 donations by the first-generation test (HCV 1.0) and 915,000 donations by the second-generation test (HCV 2.0). Donations with positive EIA results were confirmed with a recombinant immunoblot assay. RESULTS: Of these donors, 1,637 (0.13%) were confirmed as HCV-positive and 21,666 (1.72%) had elevated ALT. To estimate the additional margin of safety due to ALT screening, all donors who seroconverted were reviewed, and those donors who had elevated ALT but were HCV negative on a previous donation were identified. One hundred eleven HCV seroconversions were observed: 19 seroconversions from HCV 1.0-negative to HCV 1.0-confirmed-positive, 82 apparent seroconversions from HCV 1.0-negative to HCV 2.0-confirmed-positive, and 10 seroconversions from HCV 2.0-negative to HCV 2.0-confirmed-positive. The number of apparent HCV 1.0-negative to HCV 2.0-positive seroconversions was much greater than expected, which reflected the increased sensitivity of HCV 2.0. Only 15 donors were identified who had an elevated ALT on a previous HCV-negative blood donation, and all of these were among those who apparently seroconverted from HCV 1.0-negative to HCV 2.0-confirmed-positive. Out of the 10 HCV 2.0-seroconverting donors, no donor was found who was initially HCV 2.0 negative with elevated ALT and later was HCV 2.0 positive; nor were such donors found among 4 additional HCV 2.0-seroconverting donors. CONCLUSION: With the introduction of HCV 2.0 screening. ALT appears to have little value as a surrogate test for hepatitis C, and ALT testing was unable to detect any donors who later seroconverted, as detected by HCV 2.0.

Transfusion-associated circulatory overload in orthopedic surgery patients: a multi-institutional study.

Although recognized as a serious complication of hemotherapy, few data are available on the incidence of transfusion-associated circulatory overload (TACO). Detailed demographic and clinical information was obtained from records of 382 Medicare patients undergoing total hip or knee replacements (and receiving transfusions) from January 1992 to December 1993 at five Massachusetts hospitals. Seventy-eight percent of the patients were women with a mean age of 77 years. Thirty-two percent had co-morbidities including myocardial or coronary disease. Transfusion-related complications and comorbidities were identified and reviewed by transfusion experts. Patients were excluded from consideration if non-transfusion factors such as myocardial disease could have contributed to the development of acute pulmonary edema. Four (3 females, 1 male) patients (1.05%) developed TACO postoperatively. Mean age of these patients was 84 years (range, 75-101) versus 77 years for non-TACO. The mean intraoperative estimated blood loss was 375 mL. Each patient received only 1-2 units of red blood cells prior to onset of TACO, and in two cases only autologous blood was used. The mean positive fluid balance was 2,480 mL. The mean pretransfusion hematocrit prior to circulatory overload (CO) was 26.0 percent. Symptoms were reversed with diuretics. Length of stay was significantly prolonged by these incidents. TACO is a frequent and serious event in an orthopedic surgical setting. It is associated with advanced age, increased health care costs, and may occur in the setting of modest transfusion volumes. The utilization of conservative transfusion criteria and fluid management in the perioperative setting may decrease the incidence of this complication in this population.

Characterization of reactions after transfusion of cellular blood components that are white cell reduced before storage.

BACKGROUND: During the storage of cellular components before transfusion, cytokines that may mediate transfusion reactions are released from white cells (WBCs). Adverse effects of transfused cellular blood components therefore depend not only on the number of residual WBCs in blood components, but also on the timing of WBC reduction. STUDY DESIGN AND METHODS: Febrile nonhemolytic transfusion reactions (FNHTRs), allergic reactions, and other reactions were characterized in recipients of 4728 units of red cells (RBCs) and 3405 bags of single-donor apheresis platelets (SDAPs), all of which underwent prestorage WBC reduction. To delineate the impact of prestorage versus poststorage WBC reduction of RBCs on transfusion reactions, these results were compared with reactions occurring after the transfusion to similar recipients of 6447 bags of RBCs that underwent poststorage WBC reduction by bedside filtration and 5197 units of SDAPs that underwent prestorage WBC reduction. The levels of interleukin (IL) 1 beta, IL-6, IL-8, and tumor necrosis factor-alpha (TNF-alpha) were measured in a subset of 20 implicated cellular blood components at the time of transfusion reactions and correlated with the duration of storage before transfusion. RESULTS: The incidence of reactions was greater after transfusions of SDAPs (5.49%) than of RBCs (1.63%). The incidence of FNHTRs after transfusion of RBCs that were WBC reduced before storage (1.1%) was significantly lower (p = 0.0045) than that after transfusion of RBCs that were WBC reduced after storage (2.15%). Although allergic reactions to RBCs that were WBC reduced before storage were also less common (0.41%) than those to RBCs that were WBC reduced after storage (0.51%), the difference was not significant (p = 0.067). At the time of reactions to RBCs and SDAPs that were reduced before storage, the level of IL-6 was negatively correlated (r = -0.54, p = 0.014) with the duration of storage before transfusion, and there was no correlation between the level of either IL-1 beta or IL-8 and the interval before transfusion. TNF-alpha was not detectable in any implicated component. CONCLUSION: FNHTRs, but not allergic reactions, were less common after transfusion of RBCs that were WBC reduced before storage than after the transfusion of those WBC reduced after storage at the bedside by filtration. The level of IL-6 in implicated cellular blood components that were WBC reduced before storage was inversely correlated with the length of storage before transfusion. Further studies are needed to determine whether the transfusion of cellular blood components that were WBC reduced before storage can both diminish the incidence of adverse reactions and improve outcome.

Transfusion practices in human immunodeficiency virus-infected patients.

BACKGROUND: The reported immunomodulatory effects of transfusion raise concern about the potential for virus activation and tumor growth in human immunodeficiency virus (HIV)-infected patients. In the absence of "standards" of transfusion practice for such patients, a survey of transfusion policies among institutions specializing in the care of HIV-infected patients was performed to delineate current practices. STUDY DESIGN AND METHODS: A survey developed by the Transfusion Practices Committee of the American Association of Blood Banks was sent to 47 AIDS clinical trial units and 14 regional hemophilia centers in North America. RESULTS: Forty-three percent of centers completed the survey. Most centers observed more than 200 HIV-infected patients each. The key findings were that 1) 81 percent of centers used identical red cell transfusion criteria for HIV-infected and noninfected patients; 2) 52 percent used recombinant human erythropoietin as initial treatment for zidovudine-induced anemia, while 46 percent used recombinant human erythropoietin for anemia not associated with zidovudine; 3) 35 percent of centers used white cell-reduced blood components in lieu of cytomegalovirus (CMV)-seronegative components when administering transfusion(s) to CMV-seronegative patients; 4) 27 percent gamma-radiated cellular components, but no case of graft-versus-host disease had been observed; 5) > 85 percent of centers used monoclonal factor VIII for pediatric and adult hemophiliacs infected with HIV; 6) approximately one-third of centers routinely white cell-reduced cellular components; and 7) the most common reasons for white cell reduction included reduction of febrile reactions and CMV risk, reduction of platelet alloimmunization, and delay of immunomodulatory consequences of transfusion. CONCLUSION: There is marked heterogeneity in transfusion practice for HIV-infected patients. Modification of cellular components to achieve different objectives is routine in many centers.

Donor follow-up of influenza vaccine-related multiple viral enzyme immunoassay reactivity.

False-positive enzyme immunoassay (EIA) tests in blood donors receiving influenza vaccine were first reported in 1991. We conducted follow-up testing for 6 months of those donors with multiply reactive, but unconfirmed EIA (at least 2 positives in anti-HCV-1.0, anti-HIV-1, and anti-HTLV-I assays) with a history of recent flu vaccine to determine the duration of false positivity. Of 133,000 donors tested, 16 met study criteria; all 16 were reactive for anti-HCV, 10 were reactive for anti-HIV-1, and 12 were reactive for anti-HTLV-I. Fifteen donors were available for follow-up testing (using the original screening and supplemental tests): 10 (67%) reverted to negative for the 3 tests and 5 remained false positive for various markers at last sampling (3-6 months after vaccination). The mean duration of false positivity for those reverting to negative EIA test status, was 4.2 months (range 2-7 months) indicating a transient phenomenon and supporting studies which suggest a role for IgM in the mechanism.

Computer interview for screening blood donors for risk of HIV transmission.

To test the ability of a computer-based interview to detect factors related to the risk of the human immunodeficiency virus (HIV) among potential blood donors, and to determine donor reactions to the use of the computer, we compared the rate of detection of HIV-related factors elicited by the computer interview with the rate elicited by standard American Red Cross procedures (written questionnaires and face-to-face interviews) for assessment of donor suitability. The study was performed at a Red Cross blood donor center and a hospital. A consecutive sample of 294 male and female blood donors 18 to 75 years of age participated in a randomized crossover trial in which the order of the two methods was reversed. Among 272 prospective donors who provided complete data, the computer identified 12 who reported either behavior associated with a risk of acquiring HIV or symptoms compatible with AIDS. None of these 12 was so identified by face-to-face interviews or written questionnaires. Only one used the confidential unit exclusion procedure to prevent use of his donated blood. Tests for antibody to HIV were negative in blood from all 272 subjects. The subjects enjoyed the computer interview and judged it to be more private than the standard method for donor assessment.