Is the routine crossmatch obsolete?

This hospital has encountered no problems since dropping crossmatches for most patients who may require transfusion. It

relies on other, simpler procedures instead.

One year ago this month, the transfusion service at University Medical Center of Southern Nevada discontinued routine crossmatching of red cell products-principally packed red blood cells but also whole blood and washed RBCs.

Pretransfusion compatibility testing continues, but it includes a major crossmatch (a serologic reaction between intended recipient serum and donor cells) only if one of the following conditions is met: 1) the patient's atypical antibody screen is positive; 2) the transfusion service's card file indicates a history of an antibody or a prior transfusion reaction; or 3) the patient has a positive direct Coombs test. When a major crossmatch is performed under such circumstances, it is a three-phase procedure-immediate spin, low-ion-strength saline solution at 37 C for 15 minutes, and an anti-IgG antiglobulin phase.

The current procedure for patients who do not require crossmatches-as described above-includes a check of previous transfusion service records, an ABO and Rh performed by two different technologists, and a three-cell atypical antibody screen. ABO- and Rh-identical or compatible blood is then dispensed to the ward clerk or nurse directly off the shelf. Appropriate labels indicating compatibility are affixed at the time the blood is released.

To some readers, this protocol may seem a significant departure from standard transfusion practice, but it is just a small step in a logical evolutionary process. It is medically sound, cost-effective, and consistent with current standards of the American Association of Blood Banks.

Far from being monolithic, what is called a "crossmatch" has seen many changes and variations since the turn of the century.' From the late 1940s through the late 1960s, blood bankers tended to add steps to the basic major crossmatch on the assumption that such additions would improve its sensitivity and, ultimately, transfusion safety.

At the height of this activity, major and minor crossmatches were performed along with room-temperature incubations, enzyme crossmatches, and during the 1960s even interunit crossmatches (to be certain that atypical antibodies in one unit would not agglutinate or hemolyze cells in another unit that was to be transfused during the same time frame).

In about 1970 blood bankers began to realize their procedures were becoming too cumbersome, and they questioned whether such elaborate testing increased or decreased transfusion safety. The doubts reached a crescendo toward the end of the decade. They resulted in a reduction or elimination of a number of practices, or at least moves to make the practices optional: the minor crossmatch, the enzyme crossmatch, the room-temperature incubation, the mandatory use of broad-spectrum Coombs serum-as opposed to monospecific anti-IgG-and eventually the Coombs crossmatch and the immediate-spin crossmatch.

An article published in February 1980 by Paul J. Schmidt, M.D., provides perhaps the most dramatic example of how medically irrelevant crossmatch procedures had become by the end of the 1970s. Dr. Schmidt summarized data filed with the Food and Drug Administration on the first 69 transfusion-associated fatalities since reporting of such fatalities became mandatory in December 1975. The data clearly demonstrated that the patients had died because of clerical errors resulting in ABO-incompatible transfusions.. Exotic antibodies were never the cause. It was obvious that we needed to direct more effort against ABO errors instead of spending so much time worrying about Xg sup a and Le sup b.

As part of the reevaluation of the crossmatch, the routine Coombs phase eventually became optional. The underlying logic for this dated back to 1964, when Morten Grove-Rasmussen, M.D., suggested that the Coombs crossmatch was superfluous following negative results from a carefully performed atypical antibody screen.

Although the routine Coombs phase was demoted to optional status, the immediate spin was retained. The thinking was this would detect an ABO error that might lead, for example, to a group 0 patient receiving group A blood.

The immediate-spin crossmatch was retained for another reason as well: so that blood bankers would not alarm their clinical colleagues by telling them "the crossmatch" (as opposed to pretransfusion compatibility testing) was no longer required.

Even though the crossmatch was reduced to a vestige of its former self in those institutions that dropped the routine Coombs phase, clinicians still got what they wanted-serologically safe blood with a label that said so. We bring this up because a crossmatch slip indicating "compatibility" is an important guarantee for clinicians who are unfamiliar with how compatibility testing is performed and interpreted. They understand that when the laboratory goes on record stating that blood is compatible, it has put its credibility on the line.

Consumers react in the same manner to a Good Housekeeping Seal of Approval or a sticker from Underwriters' Laboratory. We don't care how our electric blanket was manufactured and tested; we just want to see a sticker attached that tells us an expert is on record that the product in question is safe.

We belabor this point because, though a routine "crossmatch" is no longer performed at University Medical Center, we continue to do routine "pretransfusion compatibility testing." Our labels so indicate, and clinicians find this acceptable.

In the spring of 1987, Ira A. Shulman, M.D., and his associates at Los Angeles County/ University of Southern California Medical Center suggested that the immediate-spin crossmatch was not the best way to detect ABO errors. They proposed supplanting it with a "direct ABO verification test." In other words, perform the ABO grouping on the intended recipient a second time and dispense with crossmatching entirely. (This would pertain providing the intended recipient has a negative atypical antibody screen and direct Coombs test and a negative transfusion reaction history.)

Their data indicated that an ABO-incompatible unit might be missed in one out of 500 immediate-spin crossmatches. More significant, when an AB unit of blood was crossmatched by immediate spin against group A recipient specimens, the false-negative rate was 3.7 per cent; against group B recipient specimens, the false-negative rate was 7.7 per cent. (W. John Judd and associates recently explored the mechanism of such false-negative immediate-spin crossmatches. They showed that, in addition to the well-known phenomenon of low-titer anti-A and/or anti-B among recipients, false-negative immediate-spin crossmatches may be a consequence of prozone effects.)

The Shulman procedure seemed like a good idea to us at University Medical Center for several reasons. First, we had already dropped the routine Coombs crossmatch several years earlier. Second, since 1970 University Medical Center (then Southern Nevada Memorial Hospital) had been performing ABO and Rh verification by a second technologist, out of concern that an ABO error was the most dangerous serologic event in transfusion. So we were already performing the very procedure that the Los Angeles team advocated-plus an immediate-spin crossmatch.

Third, our atypical antibody screen used a three-cell panel - instead of a two-cell panel to maximize its sensitivity since we no longer relied on a routine Coombs crossmatch to detect atypical antibodies.

After several discussions among the pathologists and laboratory supervisors, we decided to implement the procedure suggested by Shulman et al. We believed it would save time and money without compromising patient care-important considerations for all public hospitals faced with tight budgets.

Since our transfusion service is accredited by the AABB, we first had to make sure the new procedure complied with AABB Standards. Through discussions with Dr. Shulman, the primary author of the Los Angeles report, and subsequent correspondence with the AABB Standards committee, we were assured that the proposed protocol was in compliance with section G3.200 of the 1987 Standards. Thus it did not require AABB approval as a deviation from standards or as a "private standard."

Our transfusion service is also accredited by the College of American Pathologists, which recently approved this protocol.

We next presented the concept to the laboratory committee, the department of surgery, and the medical executive committee. All three agreed with our reasoning and endorsed the plan, but some anxiety was expressed. To ease this anxiety, the laboratory agreed to a continuous review of any problems that might arise, together with the hospital's quality assurance department. No problems have arisen so far.

Implemented on May 1, 1987, the new protocol soon proved itself medically while reducing the time and paperwork required to provide services. As a result, clinicians' concerns gradually abated. The procedure was "transparent" to clinicians-from their perspective, there was no change in protocol.

Figures I and II show the volume of transfusion activity at our blood bank for most of last year and the number of patients who required traditional crossmatches since the change in protocol. Reasons that patients required crossmatches are listed in Figure III.

Here are some further advantages of the new protocol to the transfusion service:

* Units remain refrigerated until ready to be dispensed,

*There is no loss of "pigtails" from repetitive crossmatching.

*Much fewer compatibility slips have to be made out. Only 46 per cent of the crossmatched units were transfused in the past, but the slips had to be attached to 100 per cent. After 48 hours, the slips were removed from unused units.

MA copy of the compatibility slip, stating "blood crossmatched but not transfused," used to be placed on the patient chart. This is no longer necessary.

*Billing is simplified for pretransfusion testing.

*It is much faster and easier to locate a correct unit in the blood bank refrigerator.

*Clerical checking of crossmatch slips versus blood units transfused has decreased by 54 per cent.

*It is easier to control short-dated units so that they are not overlooked and don't expire.

Our protocol continues to provide serologically safe blood for transfusion at significantly lower levels of CAP workload recording units. We feel it is a rational alternative to other compatibility testing protocols. And we hope a new lab computer system to be employed in the future will further enhance transfusion safety by automatically locking out the release of ABO-incompatible units.