Blood grouping prerequisites for electronic crossmatching in a fully automated environment (including electronic transfer of results)

A colleague in New Zealand wonders if, for pre-transfusion testing of previously unknown patients in a fully automated environment (including electronic transfer of results), it is considered acceptable to have the analyzer perform the initial and confirmatory blood group testing on the same cell suspension. She wonders if anyone can give references to any national/international guidelines or standards that state a requirement for the initial and confirmatory blood group testing to be performed on separate cell suspensions, in the above scenario.

The following comments have been received.

ADDENDA May 4, 2005

1. A colleague in The Netherlands is of the opinion that the chance of a laboratory error during pre-transfusion testing in a fully automated environment is probably minimal, but this approach will not detect all phlebotomy errors. The use of electronic patient identification systems will reduce these errors. He adds that in The Netherlands it is required that pre-transfusion testing of previously unknown patients include an initial sample for ABO-Rh(D) typing, and that a second sample must be obtained at a different time to be ABO-Rh(D) typed. The results from the first sample are compared to the results from the second sample. The results must agree prior to issue of blood for transfusion. In case of ABO-Rh(D) discrepancies both ABO-Rh(D) results are discarded and two new blood samples are to be ABO-Rh(D) typed. In emergency situations, until a first blood sample has been ABO-Rh(D) typed, only group O-Rh(D) RBCs are issued to all patients. In female patients, group O-Rh(D) neg RBCs and in male patients group O Rh(D)-compatible RBCs are issued until the second sample has been tested.
   
   
2. A colleague from Australia believes that the two references provided are germane to the present discussion.
   • Guidelines for compatibility procedures in blood transfusion laboratories: Working Party of the British Committee for Standards in Haematology Blood Transfusion Task Force. Transfusion Medicine, 2004.
   • Guidelines for blood bank computing: Prepared by a working party of the British Committee for Standards in Haematology Blood Transfusion Task Force Working Party.Transfusion Medicine 2000.
   
   
3. A colleague in the UK reports that in his interpretation of the British Committee for Standards in Haematology (BCSH) guidelines from the UK, as they currently stand, they are not explicitly clear on this matter but do lead one to favor the option of requiring two separate aliquots from the same sample. He adds that the current computer guidelines simply allow under the conditions specified by the original question that the ABO/RhD group of the patient may be determined twice using the 'same sample'. However, if you read the guidelines on compatibility procedures in blood transfusion laboratories published in 2004, under grouping anomolies it suggests 'repeats should be performed using washed cells. To prevent the perpetuation of mistakes the cells used should be taken from the original sample rather than from a suspension made previously'. The new computer guidelines that are in final draft to be published later this year currently state 'Provided that sample handling and identification are fully automated and results are transferred electronically with no manual editing, the ABO/D group of the patient may be determined twice using two aliquots from the same sample.' Assuming that this remains in the finished guideline, this will make the guidance explicit in this regard.

ADDENDA May 8, 2005

4. Suzanne H. Butch, MA, MT(ASCP)SBB of the University of Michigan (attribution used with permission) is of the opinion that having two blood types on file in order to perform an electronic crossmatch was designed to prevent errors within the laboratory. When initially designing her own process for the electronic crossmatch, she and her Michigan colleagues performed all of ABO and Rh testing manually and entering the result of each tube directly into the Laboratory Information System (LIS). The LIS then calculated a blood type based on tables. They progressed to performing one ABO/Rh on an automated typing instrument combined with a manual tube confirmation if there was no existing type on file for the patient. This was due to a lack of an instrument/LIS interface and the automated results were entered from an instrument printout into the LIS manually. As they have again re-designed their process for an instrument/LIS interface, they continue to want to perform two typings. These typings would be performed in two separate instrument runs. This would ensure a separate identification and aliquoting of the sample. Without this separation of identification and aliquoting of the specimen, they are not sure what is gained by doing two types. It would seem to be measuring the reproducibility of the test system, not confirming a patient’s type. Depending on the LIS some hospitals, this might create a computer problem. Doing the same test in the same run might cause a conflict because in the LIS, the ABO/Rh verification step would be processing two results from the same patient at the same time. The two results would be hitting the same patient file for verification of the current typing with previous typing. The system may have been designed to prevent a “user conflict”. Suzanne adds that there are advocates for the two ABO/Rh types to be performed on samples collected at different times. This proposal implies that we cannot trust specimen collection. If we cannot trust blood bank specimen collections, can we trust specimens collected for other laboratory tests? If not, then all laboratory tests should be done twice on separately collected samples. Mislabeled specimens, where the blood in the tube is not the blood of the patient identified on the label of the tube, is a dangerous sample in the clinical laboratory in general. A critical glucose or potassium value that is treated unnecessarily has the potential to cause significant harm to a patient. She further comments that we need to get at the root cause of specimen mislabeling, and we need to design/implement systems that are easy to use and limit the risks of getting the wrong patient’s blood in the tube. She feels it is wrong to single out pretransfusion testing specimens, because that implies that it is OK to mislabel other specimens. She believes that we need to label blood specimens the same way every time using the safest system possible for all specimens submitted for laboratory testing.

ADDENDA May 10, 2005

5. W. John Judd, FIBMS, MIBiol (Professor of Immunohematology at the University of Michigan - attribution used with permission) asks some rhetorical questions about the need for more than one specimen being tested for ABO before using an electronic crossmatch test: 1) "If you use a validated instrument, with validated positive sample identification, validated interfaces and validated interpretation tables, wouldn't repeat testing of the same sample for ABO/Rh be redundant? 2) If repeat testing for ABO/Rh is not redundant, does that mean that our validations must be prone to error!!??? All kidding aside, John asks a very important question which gets to the heart of the issue, namely, if it is required to test two independently collected samples as a requirement for the electronic crossmatch, shouldn't such a rule be applied to all types of crossmatching, and not just the electronic crossmatch?

Submit comments to the e-Network Forum at enetworkforum@cbbsweb.org

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

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

Elizabeth M. St. Lezin, MD
CBBS e-Network Forum Associate Editor & Moderator