Is a computer crossmatch in the absence of an immediate-spin antibody screen adequate for persons identified to be at increased risk of forming new blood group antibodies?

A Canadian colleague is curious to know how others view the safety of using a computer crossmatch under the circumstances described by Sandler, Abedalthagafi and Langeberg in their recent article in Transfusion "Is a computer crossmatch in the absence of an immediate-spin antibody screen adequate for persons identified to be at increased risk of forming new blood group antibodies? Transfusion 2008 Oct;48(10):2265-6. In that article the authors report a case that illustrates how patients with newly formed alloantibodies may be matched and potentially transfused with incompatible RBCs when using an automated solid-phase antibody screen followed by a computer crossmatch. They describe a patient who was transfused 35-days ago who now had a negative antibody screen by automated testing but reacted strongly by immediate spin crossmatch. Anti-Fya was subsequently identified and presumably it was newly formed from the recent transfusion. If they had used a computer crossmatch, serologically incompatible blood would have been transfused with unknown consequences. The AABB Standards (25th edition) requires that blood samples for routine compatibility testing be collected within 3 days of the transfusion if the intended recipient has been transfused or pregnant within the past 3 months (Standard 5.13.3.2). In light of the potential for an automated screen paired with a computer crossmatch to miss some newly formed antibodies, Sandler, Abedalthagafi and Langeberg propose:

1. a 3-month delay for using a computer crossmatch for patients transfused or pregnant with the past 3 months if compatibility test does not include an immediate spin or other room temperature phase in the antibody screen;]
2. during the delay period, release of RBC after a negative immediate-spin crossmatch.

The authors propose a significant change based on a single case at their institution, and what is surely a rare event, i.e., a saline-reactive antibody with a clinically significant specificity non-reactive by automated solid phase, and whose clinical consequences are admittedly unknown.

How do others feel about the recommendation by Sandler, Abedalthagafi and Langeberg?

The following comments were submitted.

ADDENDA April 5, 2009

1. Suzanne Butch, Administrative Manager of the Blood Bank and Transfusion Service of the University of Michigan Hospitals in Ann Arbor, Michigan (attribution used with permission) writes:
   
   In 1979, the AABB published a book edited by Malcolm Beck and I titled "Clinically Significant and Insignificant Antibodies". In the chapter titled "Antibodies Reactive at 30 Centigrade, Room Temperature, and Below", Peter Issit wrote "There is no doubt that antibodies whose total reactions are confined to temperatures below 30C are completely benign in vivo". He further states "Recently, the room temperature incubation portion of the test (antibody screen and crossmatch) has become the subject of critical review and those who doubt its worth have suggested that it detects only antibodies that would better not be found at all."
   
   The AABB Standard 5.13.3 indicates that we need to test for unexpected antibodies to red cell antigens using "methods of testing.. that demonstrate clinically significant antibodies. They shall include incubation at 37C preceding an antigliobulin testing using reagent red cells that are not pooled." Since neither an immediate spin reading nor an incubation at room temperature is required as part of the antibody screen or crossmatch, I conclude that antibodies reactive only in those conditions may not be clinically significant.
   
   Since most of us would rather find the positive screen than the positive crossmatch, an alternative conclusion could be that the sensitivity (without sacrificing specificity) of the antibody screening method should be improved.
   
   We have performed over 900,000 electronic crossmatches since 1992. We have yet to find an instance where, in retrospect, a serious reaction has caused us to rethink our process and revert back to a serologic crossmatch.
   
   We also know that no one method of antibody detection or crossmatching will find all antibodies that are potentially clinically significant. Beyond meeting the AABB Standards, the choice of methods for antibody detection and crossmatching is up to the facility. We cannot do the experiments we might like to resolve the question of the clinical significance of some antibodies. Was the Anti-Fya clinically significant? If it was clinically significant, and was transfused, would the transfusion still have provided adequate treatment to the patient even if cells were removed a few days earlier from the patient's blood stream? All are interesting academic questions.

2. The Transfusion Service Supervisor at a Mid-Atlantic general hospital wonders "how did this laboratory discover the saline reactive antibody since the screen was negative by their automated method and they do not do an immediate spin crossmatch? Did this antibody become known because of a transfusion reaction?" She adds that p. 919 and 920 of Applied Blood Group Serology, 4th ed. suggest it wouldn't make any difference except for a very rare instance, but that even compatible crossmatches can be incompatible.

3. The Lead Clinical Laboratory Scientist for a well-known Medical Center in the California Desert writes:
   
   I think I would be more concerned with why the antibody wasn't detected by the solid phase screening method. We had an Anti-Fya that appeared to be IgG-4 subclass since it reacted only with Anti-IgG reagent from Ortho Diagnostics and not with Anti-IgG from Immucor (Immocor states in their package instert that pure IgG4 subclass antibodies will not be detected with their reagent).
   
   Changing a protocol based on one rare case would be like doing all antiglobulin crossmatches again because you might detect incompatibility due to a low incidence antigen not found on the antibody screen.
   
   Logistically, how would you know if the patient had been pregnant or transfused within the last three months unless you had a method to query every possible transfusion receipient?

Submit Comments to the e-Network Forum below:

Items marked with an asterisk (*) are required entries. Please note that there is a 60 minute time-out on this form. if one anticipates taking more than 30 minutes or so to compose the submission, we recommend composing your comments in a word processing application, and performing a copy and paste of your text into the comments field.
* Name
Title
Affiliation
* email address
* Enter email address again to verify
* Do you wish to have attribution for your comments?	Yes No
* Your Comments
*Please indicate if this is a new question, or if you are commenting on an existing discussion:	New Question Existing, Name:
File Attachments (2MB limit)
To help us prevent SPAM, you must enter the the validation code EXACTLY AS IT IS DISPLAYED in the image directly to the right of the text entry field below in order to submit this form. Please note that the validation code is CASE SENSITIVE.

Ira A. Shulman, MD
CBBS e-Network Forum Editor & Moderator

W. Tait Stevens, MD
CBBS e-Network Forum Assistant Editor & Moderator