Feature Story

Perks plus: the new hematology analyzers

June 2002
William Check, PhD

Steven Kroft, MD, conducted a search for a new hematology analyzer in the past year. His short list of mandatory features included substantially decreasing the manual review rate for differentials while automatically quantitating nucleated red blood cells, reticulocytes, and platelets. "Pretty much all the high-end instruments we looked at filled these requirements," says Dr. Kroft, assistant professor of pathology in the Division of Hematopathology and Immunology at the University of Texas Southwestern Medical School in Dallas.

Dr. Kroft found that, beyond satisfying his basic criteria, several instruments offered additional novel and desirable applications. Some are FDA approved, such as early identification of anemias and measuring the immature reticulocyte fraction. But the more exciting advances are still investigational—enumerating immature granulocytes, estimating stem cell numbers by quantitating surrogate cells, counting reticulated platelets, and handling non-traditional specimens, particularly bone marrow. Moreover, some instruments have the potential to offer further novel and advantageous features, such as immunological assays, with only software modifications. This enhancement was one factor in breaking the tie between the top two instruments on Dr. Kroft’s list.

Also important was the ability to be incorporated into what Dr. Kroft calls "a hematology island of automation"—analyzers integrated with an automatic slidemaker/stainer under a central controlling computer that coordinates the movement of samples. In this system, which Dr. Kroft’s laboratory is gradually implementing, he says, "If we needed to repeat a CBC for some reason, no one would have to handle that tube. It would be done automatically."

Ed Wong, MD, director of hematology in the Department of Laboratory Medicine at Children’s National Medical Center in Washington, DC, also recently purchased a new hematology analyzer. "The more modern instrumentation available today has improved on original designs," Dr. Wong says, "and we wanted to take advantage of that to see if we could improve patient care." Many contemporary instruments "offer a lot in improved white cell, red cell, and platelet analysis," he notes.

Dr. Wong is satisfied with the instrument he chose, the Roche Sysmex XE 2100. "We now have an automated machine that can do everything we want in terms of reticulocyte and platelet counts," he says. It also satisfies the needs of his medical center’s bone marrow transplantation and hematology-oncology programs.

A major technological advance—incorporating a laser into the instrument’s standard operation-has eased automated hematology analyzers from the shadow of unfulfilled promise into the bright light of accomplishment. (An argon ion laser is used in all instruments except the Sysmex, which contains a semiconductor diode laser.) "More of the current analyzers are moving in that direction," says Dr. Kroft, who chose the Abbott Cell-Dyn 4000. "They are essentially little automated flow cytometers for the hematology laboratory." But instruments differ in the degree to which they "exploit that technology to do traditional flow cytometric assays," he notes. One of the factors in his choice of the Cell-Dyn 4000 was its greater ability to do flow cytometric assays, particularly CD4/8 subsets, offering considerable labor savings. "Anything that is labor-intensive and high volume is fair game for this type of analysis," Dr. Kroft says.

Carol Briggs, head of hematology evaluation and development in the Department of Haematology at University College London Hospital, sees, too, the trend that Dr. Kroft describes. "Basically all vendors are moving towards using more flow cytometric methods on hematology analyzers, combining the research side with the routine side," she says. "It would be lovely to be able to measure CD4 and CD8 counts on HIV patients straight through on your blood sample," she says, "and I think that will be coming." Another upcoming application will be measurement of immature platelets, which, she says, are getting more important. "As research techniques become more established, people want them done routinely and with one tube of blood and reduced number of staff and quickest turnaround," she says. A hematology analyzer with a fully exploited laser could meet those goals. She sees CD4 and CD8 counts as an obvious first step; the Cell-Dyn 4000 is the only analyzer now that measures them, though not at the same time as the CBC. "A separate tube of reagent is needed," she says, "and the test has to be preselected."

Bruce Davis, MD, associate director of clinical research at the Maine Medical Center Research Institute, says that several high-end hematology instruments have a limited capability to perform many of the same measurements that are on multi-parameter flow cytometric instruments, "given that they use a combination of laser excitation and fluorescence." The crucial question determining the extent to which tests will move from flow cytometry to hematology is, "What would be the volume, and is the cost of development justified by clinical utility," says Dr. Davis, who is also medical director of flow cytometry at Genzyme Genetics in Santa Fe, NM. "Vendors must balance cost of development against return on investment."

Going from the long-term prospect of doing flow cytometric assays on a hematology instrument to the more mundane hope of reducing manual differentials is like driving away from a Ferrari showroom in your Volvo station wagon. For anyone running a hematology laboratory, however, a truly automated differential would be as welcome as cost-effective, reliable daily transportation.

"One of our big issues was that we were bottlenecking at slide review, which was keeping our technologists very busy and increasing our turnaround times," says Dr. Kroft. "We have a very high smear review rate at Parkland [Hospital], in part because we have many neonates and extremely sick patients." A major issue for Dr. Kroft in selecting a new analyzer was to maximize the sensitivity and specificity of flagging of abnormal blood specimens, which the Cell-Dyn 4000 achieves.

Dr. Wong says that the manual differential rate at his hospital ranged between 50 and 60 percent. "We think we can get it down to 35 percent or so with our new instrument," he says. Because the Sysmex XE 2100 has been online for only a few months, however, he doesn’t know what the actual figure will be.

New analyzers reduce the rate of manual differential review partly by quantitating nucleated red blood cells automatically, which makes "an enormous difference," Briggs says. "Potentially, this feature completely negates the need to do manual counts," she says. NRBCs, which are included in and distort the white blood cell count, are present in infants and in persons with hematologic disorders, such as sickle cell disease and thalassemias, as well as in some persons with cancer and infections. Older machines flagged specimens that had NRBCs, and a technologist then had to stain those samples and count NRBCs manually. Says Dr. Wong: "Newer machines can quantitate NRBCs and correct the WBC count. We have a great need for this feature in our NICU."

Dr. Kroft has a similar need. "In Parkland Hospital, there are babies around every corner," he says. Technologists were reviewing smears on every CBC from neonates because neonates usually have significant numbers of reticulated RBCs, as well as RBCs with strange morphology. Dr. Kroft calls manual review "a big time sink for our technologists." He looked for instruments that could report out WBCs without having to correct manually for NRBCs. At the time of his evaluation, only two manufacturers had a quantifiable NRBC assay that was FDA approved—Cell-Dyn 4000 and Sysmex XE 2100—so he focused on those two.

Much the same could be said of Palmetto Health Richland in Columbia, SC. "With our previous instrument, almost all of the samples from babies would give the reject code on WBCs and platelets," says laboratory supervisor Christy Knight, BS, MT(ASCP). Overcall problems were also common on sickle cell patients. When they got a Cell-Dyn 4000, Knight says, their slide reviews dropped by 50 percent.

Decreases in manual review derive also from the more-accurate automated counting of reticulocytes and platelets, which, like NRBC counting, is based on fluorescent nucleic acid dyes and laser excitation.

Reticulocyte counts, which are a fairly sensitive indicator of RBC production in bone marrow, are most commonly ordered when investigating anemias or if patients are taking certain drugs, says Bernard Fernandes, MD, head of hematopathology and the blood transfusion service at Mt. Sinai Hospital, Toronto. Reticulocyte counts also reflect recovery of bone marrow after bone marrow transplants or chemotherapy. "Manual counts have a very high degree of imprecision," says Dr. Fernandes. "I would say the manual count is of very limited value."

Previous generations of hematology analyzers based only on impedance technology were unable to perform reticulocyte counts, and the counts were therefore performed manually. They have been done on freestanding flow cytometers, which give accurate and precise results but are time-consuming. "With this generation of CBC instruments, we can get a reticulocyte count at the same time and on the same specimen as the rest of the CBC," says Dr. Fernandes, who uses a Sysmex XE 2100.

Automated reticulocyte counting is important for several reasons, says Stefanie McFadden, MT (ASCP), SH, core laboratory manager at Grant/ Riverside Hospital in Columbus, Ohio. "Probably precision is the overriding issue for me," says McFadden, whose laboratory uses a Beckman Coulter GEN*S. In her laboratory, she found coefficients of variation ranging from 11.1 percent to 5.3 percent for automated reticulocyte counting, depending on the reticulocyte concentration (from ~0.3 percent to ~5 percent of RBCs). Precision with the manual method ranged from 46 percent to 27 percent, she found.

Says Knight at Palmetto Health Richland, "We do automated reticulocyte counting and we love it." They had been doing manual reticulocyte counting because they were not able to get automated reticulocyte counting to work on their previous instrument. "We have a lot of babies who have siderocytes," Knight says. "Our previous instrument used a staining method that stained siderocytes as well, so we couldn’t do babies or patients with sickle cell disease." Since getting a Cell-Dyn 4000, she estimates that the laboratory has done two manual reticulocyte counts in the past year out of a volume of 75 to 100 samples per week.

At North Shore Long Island Jewish Health System in New Hyde Park, NY, reticulocyte counts are done on an ABX Pentra, says Teddie Owen, MT, supervisor of the Faculty Practice Laboratory. The instrument breaks down the count by hematocrit and by stages of maturity according to DNA content. Owen says that the laboratory reports out the CBC, then goes into another program to do the reticulocyte count, since "we can’t run reticulocytes at the same time as the CBC."

Platelet counting has followed the same pattern as reticulocytes: initial impedance methods, then improved accuracy and precision with optical counting in the newest analyzers. Precision at the low end of platelet counts is particularly important, Dr. Kroft says. "Low-end accuracy allows us to make more informed decisions about when to transfuse platelets," he says, citing "a significant literature" proving that optical platelet counts are typically more accurate in the low range.

University College London Hospital uses a threshold for platelet transfusion of 10 (10,000/µL or 10x10⁹/L). "Our worry with the impedance method is not knowing when the platelet count is really 10," Briggs says. As with reticulocytes, the impedance method might count RBC fragments or WBC microparticles as platelets, giving a count of 20 when it is really 10. Equally problematic, since platelets are measured by size, is that large platelets get excluded, leading to a falsely low count.

An accurate method is to stain with monoclonal antibodies against CD61 and CD41 and count on a flow cytometer. Using this value as the reference, Briggs and her coworkers found that there were problems with impedance counts at the low end and that all instruments that did optical counts were significantly better.

Says Dr. Kroft, "From reading the literature and comparing methods for automated optical platelet counting, it was apparent to me that several manufacturers now provide it on their hematology analyzers and all perform in excellent fashion." He favored the Cell-Dyn 4000 because it performs dual impedance and optical platelet counts automatically on every specimen. With the Sysmex XE 2100, for both NRBCs and platelets a qualitative screen is first done. If a flag is thrown, the sample is reflexed to generate a numeric NRBC or platelet count.

"We are currently in the process of validating that we can report out low platelet counts with confidence," Dr. Kroft says. Their lower limit before doing manual review has been 50,000/µL. With the new instrument, Dr. Kroft is comfortable dropping that to 30,000. "We will push it further, aiming for 10,000," he says.

Owen, of North Shore Long Island Jewish Health System, says she considers being able to handle the extreme low end of platelet and WBC counts a crucial requirement for a hematology instrument. "Some of our patients are very sick and many have blood cancers," she says. "Once we run our controls, that is probably the last normal we see for the day."

Beyond doing a truly automated differential count and accurately quantitating NRBCs, reticulocytes, and platelets, high-end analyzers can help diagnose anemia and reflect the hematopoietic status of the bone marrow generally. One useful value is to measure directly the hemoglobin content of reticulocytes (CHr), a patented feature of the Bayer Advia 120.

"For several years analyzers have reported mean corpuscular hemoglobin [MCH], the hemoglobin content of RBCs," says Carlo Brugnara, MD, director of the hematology laboratory at Children’s Hospital, Boston, and associate professor of pathology at Harvard Medical School. In an analogous way, CHr gives the hemoglobin content of reticulocytes in picograms per cell, calculated from direct measurements of volume and hemoglobin concentration of individual reticulocytes. Because reticulocytes reflect the functional state of the bone marrow in the preceding four to five days, CHr gives a much closer reflection than MCH of the real-time balance between iron and erythropoiesis. (Lifespan of RBCs is 120 days.)

Dr. Brugnara has evaluated CHr in a variety of conditions, including anemia of children and after therapy with recombinant human erythropoietin, or rhEpo. "There is no gold standard for iron deficiency in erythropoiesis," Dr. Brugnara says. Proving iron deficiency requires treating with IV iron and showing that patients respond with an increase in hemoglobin and CHr. In a 1997 paper, nephrologist Steven Fishbane, MD, demonstrated these responses following IV iron therapy in dialysis patients who had low CHr. "We have done similar studies in classical iron deficiency patients," Dr. Brugnara says. One advantage of using CHr is that it responds within two to four days of initiating treatment, whereas other parameters tend to improve much more slowly, with hemoglobin taking four to five weeks to increase.

A particularly damaging form of iron deficiency occurs in children ages three months to 15-18 months, when development of the brain is crucially dependent on the availability of iron. "Even with moderate iron deficiency, not severe enough to cause anemia, permanent and irreversible alterations in cognitive brain function can occur," Dr. Brugnara says. He reported in 1999 that CHr can identify patients who are iron deficient even in the absence of anemia. "We are now trying further to validate this parameter and show early iron therapy can reverse these changes," he says.

A third application of CHr is to evaluate patients receiving rhEpo. "When you use Epo, it drives the bone marrow to such a high rate of production that the iron that is available in normal subjects is not enough," Dr. Brugnara says. This creates a transient state called functional iron deficiency, which cannot be detected except with an assay that shows what is happening in bone marrow at that time. Several years ago Dr. Brugnara gave normal patients a short course of Epo; all developed functional iron deficiency and, for 10 days to two weeks, produced reticulocytes and RBCs functionally indistinguishable from those seen in patients with classical iron deficiency. Those with more iron—ferritin values ≥ 100 ng/mL—made more RBCs with the same dosage of Epo. "If you use enough IV iron to saturate the system, that allows the patient to get the maximum benefit from Epo," Dr. Brugnara suggests. With Epo, an expensive drug, being given to most of the approximately 200,000 persons in the U.S. on dialysis, maximizing its benefit is important.

Briggs has evaluated a parameter on the Sysmex XE 2100 called RET-Y, a measure of reticulocyte size obtained from the mean value of the forward scatter of reticulocytes. "We found that RET-Y correlated very well with CHr," she says, "although we didn’t go to the clinical side."

Says Dr. Brugnara: "We have shown that measurement of reticulocyte volume is not as sensitive as CHr. In patients taking Epo it changes for a variety of reasons." Most important, he says, RET-Y can’t provide the hemoglobin content per cell.

In theory, immature reticulocyte fraction, or IRF, which can be done on any analyzer that does automated reticulocyte counting, could be useful for evaluating bone marrow erythropoietic activity, Dr. Kroft notes. Reticulocyte number tells how many immature RBCs are circulating, but it doesn’t take into account their different half-lives. "With IRF," Dr. Kroft says, "for the first time we can get a sense of how fast or early the reticulocytes are getting pushed out of the marrow, the true rate of production of RBCs."

Says Dr. Davis, "At the moment, IRF is probably one of the most underutilized yet clinically beneficial parameters that high-end hematology instruments provide." Hundreds of papers demonstrate its utility, he adds. "If there is a question of whether RBC production is ongoing or adequate, IRF is the most efficacious parameter to look at. Probably its greatest utility is providing assistance in both diagnosis and therapeutic monitoring of the anemic patient."

One major branch-point in the differential diagnosis of anemia is between decreased RBC production and RBC destruction. "If production is decreased, IRF will be low," Dr. Davis says. "If production is appropriate, and the problem lies outside the bone marrow compartment, IRF is increased." Recovery of bone marrow following Epo therapy, chemotherapy, or bone marrow transplantation is reflected within a few days by increased IRF.

Immature granulocyte number can be measured on contemporary analyzers, but this parameter is not yet approved in the United States. Dr. Fernandes, in Toronto, is evaluating its utility. "Counting immature granulocytes could be useful in many clinical situations," he says. It is increased in infections, many hematologic disorders, diseases of the bone marrow, and inflammatory diseases.

Briggs notes that hematology analyzers don’t differentiate immature granulocytes into their three stages, which a slide count does. "It doesn’t matter in a patient with infection or inflammation," she says, "but for hematologic disorders—leukemia or myelodysplastic syndrome—it does."

Measuring stem cell production on a hematology analyzer is a promising but problematic proposition. Dr. Wong is using a new parameter on the Sysmex XE 2100 that measures a subpopulation, called hematopoietic progenitor cells, or HPC, that resembles or correlates with CD34-positive stem cells. "I think hematopoietic progenitor cells is a misnomer," he says. "Further work is necessary to clarify what these cells’ characteristics are." Nonetheless, he finds them useful in timing apheresis collections, because quantitation of true CD34-positive stem cells requires a flow cytometer with trained technical staff. Measuring HPC allows him to determine within 90 seconds whether to go ahead with stem cell collection.

Michael Watts, PhD, director of the stem cell laboratory in the Haematology Department at University College London, evaluated HPC. To mobilize stem cells from bone marrow into the circulation, patients are given growth factors and chemotherapy. Apheresis could be timed by doing WBC counts. However, there is a problem of "poor mobilizers," patients with extensive prior therapy or whose marrow has not yet recovered from prior treatment.

Direct measurement of CD34-positive cells in peripheral blood on a flow cytometer has been used to time apheresis. However, the concentration of CD34-positive cells in many patients’ blood is around 0.1 percent, a range in which the precision of a flow cytometer drops substantially. "Our interest in the HPC test was to find something that would reduce the number of flow cytometer measurements and improve harvest timing beyond simply WBC recovery," says Dr. Watts.

He and his coworkers showed good reproducibility for HPC even at very low levels. Correlation between HPC and CD34 number was "poor," they found, consistent with the idea that the two tests measure different subpopulations. "But that doesn’t matter since we are only looking for the recovery phase of bone marrow," Dr. Watts says. Most important, adequate apheresis yield predicted by HPC was only slightly less accurate than by CD34 counts.

An "optimal strategy" would be to proceed with apheresis for HPC counts > 10x10⁶/L (typically > 0.1 percent), with the option of doing flow cytometry for lower counts. This approach has the potential to reduce flow cytometer runs by 90 percent. For now they continue to do both tests for validation. Rare cases occur in which relying on HPC alone would result in collections with inadequate numbers of stem cells.

Dr. Davis calls the HPC count "an imperfect surrogate" and says it "does not necessarily give the count you want." Stimulation with growth factors and chemotherapy drugs drives many types of cells in the bone marrow to be produced at a higher rate, he notes. He favors the use of IRF as a surrogate to stem cell counting, based on several publications by Spanish and Australian groups. "IRF provides a parameter that parallels CD34 stem cells and several instruments already can measure it," Dr. Davis says. Dr. Kroft agrees that IRF needs to be evaluated further. "It holds promise because it is so easy to do," he says.

Dr. Davis is helping a working group of the International Society for Laboratory Hematology to develop another assay, reticulated platelets, for evaluation of thrombopoiesis. "Reticulated platelets, simply put, are the platelet analogy to the reticulated component of the RBC series," Dr. Davis says. Like reticulocytes, younger platelets—those most recently produced and released by bone marrow—have more RNA, so they are a fairly direct measurement of bone marrow platelet production.

"Measuring reticulated platelets turns out to be a fairly informative way to assess thrombopoiesis without doing more invasive techniques such as bone marrow biopsies and aspirations, which are themselves not always informative," Dr. Davis says.

Reticulated platelet measurement has been reported to discriminate between thrombocytopenias due to platelet consumption, such as hypersplenism, and those due to low platelet production. In principle, reticulated platelets could be measured on any analyzer that can do a fluorescent reticulocyte count, although no instrument offers that assay now.

All of the applications described so far use peripheral blood samples. But a few pathologists have assessed whether hematology analyzers can handle non-traditional specimens, particularly bone marrow. Harold Schumacher, MD, emeritus professor of pathology and director of the hematopathology fellowship program at the University of Cincinnati Medical Center, ran remainder material from about 300 bone marrow biopsies through a Cell-Dyn 4000 and compared the differential counts obtained with his hand counts. "I calculated the relation of the myeloid cellular elements to erythroid elements and determined abnormal cells, particularly blast cells from acute leukemia," he says. Abnormal cells are identified by position on the instrument’s multi-angle polarized scatter separation histograms, with lobularity on one axis and size on the other.

Performance of the instrument was not hampered by the fact that it was analyzing bone marrow, Dr. Schumacher found. Some people expected the presence of marrow fat to clog up the aperture. "But I never had problems with aperture occlusion," Dr. Schumacher says. In spiking experiments with leukemic blast cells from culture, the instrument proved accurate down to at least 3.6 percent blasts. A good hematopathologist can detect down to around five percent blasts on a slide, and a flow cytometer is sensitive down to one to two percent, Dr. Schumacher estimates. The next logical step would be to look at bone marrows from leukemic patients.

Maria Pedemonte, MD, medical director of Bioreference/Genpath Laboratories in Elmwood Park, NJ, is running bone marrow samples on a Bayer Advia 120. "My assumption," Dr. Pedemonte says, "was that if I ran peripheral blood or bone marrow on the analyzer, by combining the laser histogram and physical properties of the cells with myeloperoxidase staining, I would get a good idea of lymphoid versus myeloid and monocytic lineage." This was the case, she found, with some leukemias having characteristic histogram patterns. For instance, acute promyelocytic leukemia gives a very bright region in the side scatter of the myeloperoxidase-positive channel. Like Dr. Schumacher, she found no clogging from fat particles.

Dr. Pedemonte finds this method to be complementary to flow cytometry, reducing the number of antibodies needed in some cases by about one-third. "The most important thing is that this method is not diagnostic," she says. "I use it as a screening tool in conjunction with morphology, flow cytometry, and cytogenetics."

Some of today’s analyzers can also perform immunological assays, potentially moving some tests from the flow cytometer to the hematology laboratory. Says Dr. Kroft, "That is one of the things that excited us" about the Cell-Dyn 4000, the development of immunologically based assays for a hematology platform. One example is that it can do a platelet count with CD61 monoclonal antibody in an almost completely automated fashion. "It incubates a blood sample with antibody to CD61 antigen, then enumerates platelets by fluorescence measurement," Dr. Kroft says. This kind of assay done on the flow cytometer is essentially the gold standard for platelet counting. On the hematology analyzer, it can function as a backup method to confirm a platelet count that is very low or has abnormal platelets, such as very large or hypogranular platelets, as in myelodysplastic syndrome. Dr. Kroft has not yet developed the assay in his laboratory, but it is available from the manufacturer and FDA approved so he anticipates adopting it.

"That is a nice feature of the Cell-Dyn 4000," Dr. Kroft says. "It has the ability to do a variety of flow-based assays in up to three colors." One approved assay of this type is enumeration of T cell subsets (CD4 and CD8 cells) in HIV-infected patients. Because this is a relatively high-volume test, it could make economic sense to move it onto a hematology platform.

Looking further into the future, Dr. Davis sees hematology analyzers being used to measure functional states of cells, rather than just counting cells as particles. He is working on measuring expression of a protein, CD64, the high-affinity Fc receptor, on neutrophils. "I and others have a lot of evidence that neutrophils upregulate CD64 in the acute inflammatory response," he says. "So measuring CD64 levels on neutrophils could provide a good indicator of sepsis or infection." Since this method is essentially done with a monoclonal antibody and fluorescence detection, it could go on any current high-end hematology instrument.

"If there is any area where laboratory medicine has not moved forward in the last 10 to 20 years," Dr. Davis says, "it is providing new diagnostics for infection or sepsis. We still count neutrophils or look for indication of left shift or sedimentation rate as indirect measures of inflammation." He predicts that the hematology platform will bring activation markers onto a blood cell counter: "Its design is ideally suited for that."

William Check is a medical writer in Wilmette, Ill.