Feature Story

KBassay still ahead despite tough rival

Dosing Rh immune globulin

November 2003
William Check, PhD

It’s a fundamental tenet of American culture that technological innovation will be rewarded. As American philosopher Ralph Waldo Emerson said, "If a man can write a better book, preach a better sermon, or make a better mousetrap than his neighbor, though he builds his house in the woods, the world will make a beaten path to his door." But a situation arises now and then that challenges this assumption, or at least makes us question under what circumstances it might not apply.

Such a situation prevails in hematopathology with regard to evaluating fetal red cells in the maternal circulation. The traditional acid elution assay for hemoglobin F, the Kleihauer-Braun-Betke test, usually referred to simply as the Kleihauer-Betke, or KB, test, is widely acknowledged to be imprecise. A few years ago a flow cytometry assay for hemoglobin F, which is analytically superior, was introduced. Both assays are FDA-approved. Yet, in the most recent CAP HBF Survey, only 21 of the several hundred participants were using the flow cytometry assay.

Lack of a flow cytometer is often given as a reason for an institution’s not adopting the newer assay. A further consideration is that laboratory directors need to satisfy their clients who order the test, obstetricians, who frequently want stat results from this test. But neither of these obstacles is insurmountable. In fact, recent evidence shows that the results of the test are almost never used in a stat fashion.

It may be that the underlying reason why the flow cytometric method has not supplanted the KB assay is tradition combined with lack of an urgent reason to change. The KB assay was introduced in the 1940s and is familiar. Quincy Crider, PhD, president of Suretech Diagnostics, which markets a popular version of the KB test, says it has been around "forever."

"It has worked very well," he says. "Very few problems ever arose because of inaccuracy of the fetal hemoglobin assay."

Dr. Crider says that people reading KB slides tend to overcount, which means they tend to give more immune globulin than is necessary in some cases. "But that is not harmful," he says, "and it prevents underdosing, which in this case is very good."

"Change comes slowly unless there is a burning reason," says Bruce Davis, MD, president of Trillium Diagnostics, Scarborough, Me. His company developed the antibody that forms the basis of the flow cytometric assay, which is marketed by Caltag. Little attention is paid to the fetal hemoglobin assay since it is very low volume. Still, says Dr. Davis, who is also associate clinical research director of the Maine Medical Center Research Institute, Scarborough, "I have wondered, given a superior FDA-approved assay, how can you defend using an inferior assay?"

Measurement of fetal red cells in maternal blood becomes necessary when an Rh(D antigen)-negative mother is carrying an Rh-positive fetus. (The incidence of Rh incompatibility varies with ethnic group: In Caucasians, the group most affected, it makes up 15 to 20 percent of pregnancies.) Fetal hemoglobin in the maternal circulation is measured in the peripartum period to guide the use of Rh immune globulin, which is administered to the mother to protect her from being immunized against Rh antigen by fetal cells that have leaked into her blood. If immunization occurs, it can lead to complications or fetal demise in subsequent Rh-positive pregnancies. A standard dose of immune globulin is given to every Rh-negative mother who has an Rh-positive baby. In North America, this standard dose typically covers up to 30 mL of fetal hemorrhage, Dr. Davis says. An additional vial of Rhogam is given if a threshold-0.6 percent fetal cells in the maternal circulation-is exceeded. It is chiefly to determine which mothers need this additional dose that hemoglobin F is measured in cases of Rh incompatibility.

A second reason for hemoglobin F testing arises when a pregnant woman undergoes a surgical manipulation or sustains trauma, such as a car accident, or has a spontaneous abortion more than 12 weeks into the pregnancy. In this context, fetal red cells are measured to determine whether there has been fetomaternal hemorrhage due to placental injury. In these situations, testing is independent of Rh type.

"Fetal red cell testing was originally designed to dose Rh immune globulin in Rh-negative pregnant women," says William Finn, MD, clinical associate professor of pathology, director of hematopathology, and associate director of clinical laboratories at the University of Michigan Medical Center, Ann Arbor. "What has happened is that fetal red cell testing has spontaneously evolved to be a de facto measure of fetal trauma."

In any of these situations, the KB assay has drawbacks. "Historically, it has been well documented that the KB assay has both sensitivity and precision limitations," says Dr. Davis. He notes that CAP Surveys data and single-institution studies evaluating interobserver variability have revealed coefficients of variation of 50 to 100 percent for the KB assay. "In the KB test, you look through a microscope and make an estimate of color intensity on 2,000 cells," Dr. Davis says. "So it is highly subjective and has a low denominator," both of which reduce its precision.

Dr. Finn, who oversees the CAP’s HBF Survey, says two samples are sent quarterly to subscribing laboratories. Samples are made up to target values by mixing fetal red blood cells with adult blood. Laboratories are tested two ways. First they use a nonquantitative screening procedure (immune-based rosetting). A positive result on this screen triggers quantitative testing, which can be done by a manual KB-type assay or by flow cytometry.

(Some laboratories measure anti-D [anti-Rh] antibodies. "To the best of my knowledge, our Survey does not accommodate that procedure," Dr. Finn says. Anti-D applies only to an Rh-positive fetus in an Rh-negative mother for the specific purpose of dosing Rh-immune globulin; it cannot be used for direct detection of trauma or fetomaternal hemorrhage.)

Because so few laboratories use flow cytometry, calculations for that method are more difficult. "Even so," Dr. Finn says, "CVs tend to be noticeably lower in laboratories using the flow cytometric method." In the most recent Survey, CVs ranged from 36 to 54 percent for manual methods, while the CV for the flow method was 19 to 20 percent.

Surveys show not only variation among laboratories on the KB assay, but also inaccuracy classifying samples on either side of the 0.6 percent therapeutic trigger point. Dr. Davis notes that two Surveys had samples of 0.4 percent and 0.8 percent. "Results of both Surveys were amazingly similar," he says. Just under 50 percent of laboratories using the KB assay scored the 0.4 percent sample above 0.6 percent. Flow cytometry laboratories, on the other hand, did much better: In one Survey all got it right, and in the other only 12.5 percent scored it more than 0.6 percent. "Perhaps scarier from a clinical point of view," Dr. Davis says, were results on the 0.8 percent sample: Nearly 11 percent of laboratories using the KB method reported this challenge under 0.6 percent. "So there would have been underdosing in about one in 10 cases of large hemorrhage," Dr. Davis says. In both Surveys all flow cytometry laboratories got the 0.6 percent sample correct.

In the latest Survey, Dr. Finn says, there was one sample with a target value just under one percent. KB results for that sample came in above target fairly consistently, reports Dr. Finn. In fact, the mean was higher than for the flow cytometry results. "KB is less accurate but not clinically detrimental," he says.

In a recent article (Cytometry. 2002;50:285-290), three centers pooled their experience with the flow cytometry method. Of 1,248 patient samples assayed, only 69 (5.5 percent) contained detectable amounts of fetal red cells. Of those 69, only 21 patients (1.7 percent of the overall population) had fetomaternal hemorrhage greater than 30 mL. Of the 11 patients with large fetomaternal hemorrhage and clinical followup, seven (64 percent) had fetal demise. "So major hemorrhage occurs infrequently," Dr. Davis concludes, "but it can have serious implications when it does occur."

Brent Wood, MD, director of hematopathology and assistant professor of laboratory medicine at the University of Washington, Seattle, participated in this study. "We found that the flow cytometry assay works quite well," he says. It can be labor-intensive because it requires repeated washing of cells. However, Dr. Wood uses an automated cell washer of the type commonly used in blood banks. With this instrument, the flow cytometry assay works efficiently. To determine flow cytometry’s accuracy, prepared samples with a range of values were made by diluting cord blood into adult blood. Dr. Wood notes that flow cytometry’s improved precision largely comes from the fact that it typically counts 50,000 red cells, many times more than the manual assay.

Dr. Crider of Suretech attributes most problems with the KB test to lack of training in how to run the assay. "It’s not one of the big assays in blood bank or hematology," he says. "It’s a test that is actually done quite infrequently, so they don’t spend time training people adequately." Dr. Crider estimates that an average local hospital probably runs only a few KB tests per week, as opposed to ABO blood typing, which is done every day throughout the day.

"We have offered a proficiency testing program for a number of years," Dr. Crider says. "Until CAP started offering theirs, I was the only one in the world who knew how poorly this test was being run." Still, he says, "Being able to run the test well is just a matter of being trained. Overall we have seen a vast improvement in KB testing of the labs participating in our proficiency test program."

Dr. Davis agrees that a well-trained person, one who has been taught to call things the same way consistently, can reduce imprecision. "However," he reiterates, "laboratories now are not doing well. CAP HBF Surveys show continued wide disagreement or variation between what laboratories report on the same sample. In every single Survey the flow cytometry laboratories have outperformed the KB laboratories."

Suretech Diagnostics has taken a step to help ameliorate this problem. "When we became aware of how poorly this test was being run," Dr. Crider says, "we contacted Dr. Robert Davis to help us put together a training manual with photos showing how to score cells and read slides correctly." Since 1996 Suretech has offered this manual for sale to laboratories. "The photos make you aware that you have to look for more than just a dark red color," Dr. Crider says. "You have to be able to recognize cellular structure to identify fetal cells." He says the photos may not be suitable with other companies’ reagents.

Suretech also offers a complete set-negative, high, and low levels-of fetal hemoglobin controls. "Labs that are serious about reporting accurate values will find these controls helpful. They’re also a powerful training tool," Dr. Crider says. The fetal red cell controls, called Fetaltrol, are made by Trillium Diagnostics and are FDA-cleared for in vitro diagnostic use. The controls are also of help in performing the flow cytometry assay because the high fetal cell control can be used to verify that the fetal red cell counting regions are set properly.

"We have been using the Suretech kit for about eight years," says Robert Davis, MD, PhD, medical director of laboratories at Mary Rutan Hospital, Bellefontaine, Ohio. "We use it not only for fetomaternal hemorrhage at the time of delivery, but also in cases that might involve a car accident or a pregnant woman who has fallen."

Dr. Robert Davis says he finds the KB assay "extremely useful." To train medical technologists at Mary Rutan Hospital, the blood bank specialist makes dilutions of fetal cells, stains them, and does a proficiency examination for the other technologists who rotate through blood bank. "I also review the smears that he views," Dr. Robert Davis says, "and we subscribe to the Suretech check system. So we have quarterly within-department evaluations as well as external evaluations." He calls their results on both evaluations "excellent."

"We have had very good luck with the KB test clinically," Dr. Robert Davis says. "Our evaluations have shown that mothers given Rhogam based on results of the KB method have not developed antibodies, and that is the gold standard."

"Some articles state that flow cytometry is the best method, which it is," he acknowledges. "But most hospitals of our size in a rural setting do not have access to flow cytometry. We can send a specimen out, but the quicker we determine whether there is fetomaternal hemorrhage, the quicker we can administer Rhogam." He concedes there is a 72-hour window for giving Rhogam, but adds, "If you miss the courier, you have 24 hours before the sample even gets sent out. That narrows your time window. Also," he points out, "the AABB says the sooner you give Rh immune globulin, the better. The sooner Rhogam is given, the less likely the mother will become sensitized."

Dr. Bruce Davis says he frequently encounters this line of argument. However, he doesn’t find it compelling. "Certainly there is ready access to a flow cytometer within most geographic regions in North America," he says. "Or you can send samples to a reference laboratory, such as ARUP or Mayo." As for the issue of rapid turnaround, he says, "Laboratories say that physicians are expecting a stat result, even though in point of fact the way that Rhogam dosing is done there is a 72-hour window to make a dose adjustment." The first dose of immune globulin is given automatically, independent of the result of the test for fetal red cells.

Obstetricians are more likely to emphasize rapid turnaround of the fetal red cell assay in cases of maternal trauma, which is the main reason Dr. Wood does not offer this test by flow cytometry. "Sometimes it is requested at off hours, especially at night for management of emergent obstetrical patients," he says. "Probably in most cases it is not necessary to perform this test at night, but first we need to get obstetricians to accept that. Whether you can convince your obstetricians of that varies from institution to institution."

At the Mayo Clinic, James Hoyer, MD, consultant in hematopathology, uses flow cytometric measurement of red blood cells to verify a special form of hemoglobinopathy called hereditary persistence of fetal hemoglobin, or HPFH. "Certain people in adulthood have high hemoglobin F levels," he says. This condition, which is benign, might inadvertently be discovered when screening for other hemoglobinopathies, such as sickle cell disease. "The high hemoglobin F level is first seen on electrophoresis. A time-honored way of evaluating cases of suspected HPFH is to use the KB assay to determine the cellular distribution of hemoglobin F. Patients with HPFH have a pancellular pattern in which all cells contain the same amount of hemoglobin F," Dr. Hoyer says.

However, his laboratory switched from KB assay to flow cytometry. "When you are looking at the stained slide of the KB assay under a microscope for this purpose and trying to determine the hemoglobin F pattern, it’s usually very subjective. It’s much more difficult to interpret than when used in the obstetric setting," he says. "Flow cytometry gives us a very objective reading, it is a lot faster, and it allows us to store our data, which isn’t possible with the KB assay."

Like Dr. Finn, Dr. Hoyer has not been able to sell the flow cytometry method to obstetricians because they want it around the clock. "We’ve told them from our end to look at the flow cytometry method, it is very easy, but they stay with the KB assay."

Strong evidence supporting the notion that fetal red cell testing is not required, or even helpful, on a stat basis arose from a recent review at the Ohio State University Hospitals, Columbus. This study found that clinicians do not actually use the laboratory results to manage obstetrical trauma patients. The population included 2,786 pregnant women who came to the hospital in the last 17 years (1985-2002) and had a KB test performed for any reason. "In all years the majority of KB tests were done for maternal trauma, not for Rh incompatibility," says JoAnna Williams, MD, resident in pathology, who performed this review with Melanie Kennedy, MD, associate professor of pathology and director of transfusion services. Most cases were minor trauma, such as falls. For all positive and randomly selected negative tests, the investigators performed chart review to determine how the results of the KB assay correlated with clinical and fetal well-being and clinical and fetal outcome.

"If the KB assay was positive, was there any clinical indication that there was fetal hemorrhage?" Dr. Williams says. "Or was there a positive KB result and mom and fetus were fine and the mother went home and subsequently delivered a healthy baby, which we scored as a negative correlation."

They also looked at how clinicians used the results of the KB tests. Were the results even written in the chart? "A major aspect of our study was whether clinicians changed their management of patients based on the KB results. Or did they depend on clinical parameters, such as ultrasound and fetal and/or maternal heart rate?"

When all the results were tabulated, Dr. Williams says, "we found no cases out of all those positives and randomly selected negative KB results where a clinician actually changed clinical management based on the KB assay without supporting clinical data." If the KB result did not correlate with clinical data, such as a positive KB test but very reassuring maternal and fetal well-being, clinicians always managed the case according to the clinical status of the mother and fetus. "That’s a pretty dramatic finding," Dr. Williams says.

The authors concluded that the KB assay is insensitive to clinical conditions, that it has little impact on clinical care other than quantitation of RhIg dosing in Rh-negative mothers, and that there is no reason to perform this test on a stat basis since it is a poor indicator of clinical well-being. "We actually had two cases with dead pale babies with retroplacental clot confirmed by autopsy who had a negative KB assay," Dr. Williams says.

The OSU pathology department has not yet issued a formal policy change on KB testing. However, Dr. Williams says, "We have a poster of our study in the labor and delivery area, and one author who is an obstetrician has instructed all residents not to order this test on a stat basis." As a result, "there has been a very dramatic decline in stat ordering of the KB test." Flow cytometry is available in the department, but it is not yet used to assess hemoglobin F. "Our obstetricians would like us to use it," Dr. Williams says. "The problem right now is that flow cytometry is not offered on a 24-hour basis. We kind of used this study as a step to get our clinicians to accept that they don’t need this test 24 hours a day, so we could do fetal red cell measurement by flow cytometry during our normal working hours."

As for assessing Rh incompatibility, Dr. Williams says, "Stat results are not needed because you have 72 hours to administer Rhogam. If an Rh-positive baby is born to an Rh-negative mother, Rh immune globulin is effective even if given 72 hours after supposed fetomaternal hemorrhage."

Putting all these arguments and data into perspective, Dr. Finn says, "I think there has been a lot of bad press on the KB method. It is less precise than flow cytometry, but that reflects the imprecision intrinsic in manual versus automated testing."

He says he is not a "defender" of the KB test, that he is "all for flow cytometry as a better test with a lower CV and more reproducible values that are much closer to the target." But, he adds, "We need to be a bit careful here before we say that every rural laboratory needs to get a flow cytometer. Although in a perfect world everyone would have flow cytometry for this purpose, it is certainly not a practical choice if you want to deliver this service to every qualifying patient. I am hesitant to make it seem as though the College endorses one FDA-approved method over another, when our job is really proficiency testing of whatever method a given lab may choose."

Despite its less-than-perfect analytical properties, Dr. Finn says the KB assay is unlikely to be making clinical mischief. "The clinical decisionmaking algorithms build into themselves the relative imprecision of the test," he says. In his experience, most clinicians calculating Rhogam dose from fetal hemoglobin values based on KB results round up to the next vial number and then add yet another vial. "If the calculation says 1.7 vials, then you give three," he says. "In the big picture, I don’t think any patients are being harmed by application of the KB method."

A manual test is always going to be less accurate and less precise than its manual counterpart, Dr. Finn says. "But we have not abandoned manual differential counts or manual urinalysis, although we prefer to do them in automated fashion. I don’t want to come across as a defender of manual testing," he says. "I am in favor of automation. But there is a time and place for anything manual, and you have to determine that within your institution."

William Check is a medical writer in Wilmette, Ill.