The CBC at the turn of the millennium: an overview.
Since the advent of electronic cell counters, which can generate all of these test results in seconds, labor requirements have been substantially reduced. The presence of most abnormal white cells can be detected and indicated by a series of instrument-specific flags. Virtually all cases of acute leukemia can be detected by the Coulter STKS analyzer via a combination of electronic cell counting, flagging, and review of sentinel blood smears (1). However, relevant comments on red cell morphology cannot be generated by automated technology at the present time. This review will address the status of these and other issues in technology at the turn of the millennium.
Most instruments generate two types of data: graphic displays with or without flags for internal laboratory review, and a series of numbers for export to clinicians. Graphic displays take two basic forms: histograms in which relative numbers of red cells, white cells, and platelets are plotted against cell size, and scatter-plots in which white cell subpopulations are displayed. Both forms are generated by most instruments. National guidelines for evaluation of instrument performance, particularly in leukocyte differential counting, have long been established (2).
Technologies involved in earlier instruments still in use have recently been summarized (3). To these have been added some of the principles utilized in newer emerging devices. Basically, all instruments use two or more of the following methods of measuring blood cells. Impedance techniques record the momentary reduction of conductivity engendered by the passage of blood cells through tiny apertures monitored by electron sensors. Tungsten halogen and helium-neon red laser light sources also are used to analyze white cells and red cells, respectively. In the Technicon (Bayer) H-3, peroxidase histochemistry provides a method of dividing white cell populations into families. Lytic agents are used to eliminate white blood cell (WBC)  cytoplasm, leaving only bare nuclei for electronic scrutiny. The resistance of basophils to most lytic agents offers a unique approach to counting an individual cell line. The conductivity of high-frequency current is used by some instruments to determine internal physical and chemical composition of white cells. Flow technology, a feature of many instruments, is used for measurement of cell size by low-angle forward light scatter and nuclear complexity by high-angle forward scatter. This technology can be enhanced with fluorescent dyes to quantify reticulocytes and nucleated red cells. Applications of these techniques to the current crop of instruments are provided in Table 1.
Each of the available devices has its advocates from the standpoints of cost and performance. All manage to perform creditably well in counting red cells and platelets. Much of modern-day concern centers on the accuracy of WBC differential counts and on the specter of false negatives, particularly in the matter of blast detection. With improving technology and appropriate flagging, much concern regarding false negativity is evaporating. At worst, the imprecision of electronic differential counting is substantially lower than that of microscopic methods (4).
Most instruments use impedance and optical technologies to measure the total WBC count (5). Predictably, both technologies have their limitations. In impedance counting, both high and lower range accuracy may be impaired (6). Counts may also be inaccurate in the presence of circulating nucleated red cells. Optical methods, however, enjoy substantially better high and low end accuracy than impedance techniques, but they may be impaired in the presence of fragile WBCs and lyse-resistant red cells. These discordances have been rectified in the Cell Dyn 3500 analyzer [a precursor to the Cell Dyn 4000 (7) listed in Table 1] by the introduction of decision analysis based on comparisons of its dual impedance and optical modules (5).
An interesting addition to the diagnostic armamentarium of the new generation of automated analyzers is the slide-making device attached to the Coulter GEN-S. The instrument can be programmed to generate sentinel slides automatically with any red cell, white cell, or platelet flag; with location (blood labeled as originating in an oncology wing); with doctor name (oncologist); or with patient name. Sample printouts of two Coulter instruments are shown in Figs. 1 and 2.
One odd byproduct of automated technology, however, is the detectability of unlysed red cells in printouts of some cell counters (8-11). In the earlier Sysmex NE 8000, these cells were located to the left of the diagonal in the white cell scattergram. Of 13 patients studied with this instrument, 5 had previously undiagnosed hemoglobinopathies, 4 had previously diagnosed hemoglobinopathies, 1 was suspected of hemoglobinopathy, and 3 had liver disease (11). The hemoglobins detected were EE, EA, CA, SA, and SS. Common to blood smears in most of these cases were target cells, logical candidates for lyse resistance. Thus, unexpected fallout from automated technology may have identified the first poikilocyte by serendipity. Flagging indicative of the presence of unlysed red cells continues to be a feature of the newest member of the Sysmex Series (SE-9000; Fig. 3).
With the explosion of new instruments appearing on the market, many representing improvements in technology over earlier models, it is difficult if not impossible to favor a specific instrument over another at purchase time. Technologies have improved radically over the past decade. True, most systems have some limitations. However, such limitations are monitored by flags, minimizing false negatives and false positives. Although initial cost and the expense of maintenance contracts may play important roles at decision time, other factors such as amount of automation, throughput time, and patient population may be equally important. The latter to a significant degree influences frequency of flagging, a major economic consideration indeed.
Virtually all laboratories are now using automated hematology instruments to generate cell counts, red cell indices, and white cell differential counts. The addition of flags to printouts is designed to reduce the numbers of false positives and false negatives by triggering morphologic review of sentinel slides. Although this ploy has been broadly successful in avoiding interpretative debacles (1), substantial numbers of problems remain to be resolved, varying in frequency from one instrument to another. Many can be avoided at the bench by astute technologists who sequester distorted indices generated by samples taken from patients with obvious hypertriglyceridemia or cold agglutinin disease. The recent development of a new Abbott hemoglobin reagent (as used in the Cell Dyn 4000) may herald the disappearance of hypertriglyceridemia as a cause of spuriously increased hemoglobin results (7). Other error-generating aberrations, however, are silent or instrument-invisible. These include spurious macrocytosis generated by hyperglycemia (12,13) and spurious thrombocytopenia caused by platelet satellitism in EDTA blood (14). A compilation of error-generating aberrations from several sources is provided in Table 2 (12,15,16).
Although flagging serves its purpose well by indicating to the technologist or pathologist that an instrument suspects that data generated by one of its channels are marginal, the process tends to introduce its own set of problems. Automation evolved in a logical bid to reduce the labor-intensiveness of manual methods. A plethora of flags from a given instrument would therefore seem to negate the advantages of automation by precipitating large numbers of reviews of sentinel slides. The need for such reviews is predicated on several factors: (a) the gravity of the consequences of ignoring specific flags and disseminating potentially erroneous or dangerous information; (b) the establishment of tolerable limits for living with certain flags; and (c) the degree of concern of a laboratory director in mandating morphologic review.
[FIGURE 1 OMITTED]
At Northwestern Memorial Hospital (Chicago, IL), the review rate resulting from flagging approaches 30%. Such a high rate of review reflects the fact that this is a tertiary hospital, with an active oncology division, a transplant center, and a busy newborn intensive care unit. This rate is virtually identical to that reported earlier in another large university hospital (3). In contrast, the review rate of sentinel slides at St. Mary's Hospital (Duluth, MN) hovers around 5%. This 400-bed hospital, although classified as tertiary, is not a transplant center. Its low rate of sentinel slide review is influenced by two factors: (a) the large number of patients seen in a huge outpatient department; and (b) the conditioning of local physicians to order single hematologic tests rather than full complete blood counts (CBCs).
Data for Export
For most purposes, numeric data screened by the laboratory and released for consumption by clinicians appear entirely adequate in precision and accuracy. Flagging aside, the large numbers of red cells, white cells, and platelets counted in modern instruments generate much more accurate data than those deriving from older manual methods. Red cell indices are a case in point. The seven indices routinely proffered are as follows:
Red blood cells (RBC), [10.sup.6]/[micro]L
Hemoglobin (Hb), g/dL
Mean corpuscular volume (MCV), fL
Mean corpuscular hemoglobin (MCH), pg
Mean corpuscular hemoglobin concentration (MCHC), g/dL
Red cell distribution width (RDW),
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Of these, Hb (or Hct) and MCV are critical. Fortunately, both are extremely accurate. The MCHC and RDW, although also accurate, are less widely used. The RBC count and MCH are rarely used in decision-making. Without regard for possible morphologic aberrations or their nuances, clinicians by and large order appropriate second-order (follow-up) test batteries when anemias are either macrocytic (MCV [greater than or equal to] 100 fL) or microcytic (MCV <80 fL) in adults. The diagnostic possibilities in either instance are relatively limited (17). In the absence of clues from abnormal cell size, purists would consider morphologic review mandatory as a prelude to second-order testing. Although admittedly an ideal approach, such review would be less than cost-effective because of excessive labor demands. The need for morphologic review is further blunted by the distribution of anemias in the US population. Iron deficiency anemia (IDA), anemia of chronic disease (ACD), and thalassemia minor together account for >85% of such anemias and IDA/ACD for >50% (18). Furthermore, in early IDAs and in 75% of ACDs, MCVs are normal and morphologic discrimination between the two is difficult to say the least. Indices alone, notably the combination of Hb, MCV, and RDW, serve this purpose better (19). At the present time, it appears that flags rather than clinicians trigger most requests for morphologic reviews. Red cell morphology is addressed in most institutions during such reviews, even when white cell or platelet data are flagged.
Laboratory Communication with the Clinician
Routine, unflagged WBC counts and differentials are released immediately to the clinician. Flagged cell counts and WBC differentials, however, are checked morphologically by senior technologists before release. At Northwestern Memorial Hospital, WBCs <1000 or >50 000, Hb <6 or >19 (except in newborns), and platelets <50 or [greater than or equal to] 1000 000 are called directly to the clinician by the technologist. The pathologist calls the clinician directly upon encountering (a) a new hematology malignancy, e.g., circulating blasts; (b) white cells with ingested material (bacteria, fungi); and (c) schistocytes and thrombocytopenia, i.e., stigmata of disseminated intravascular coagulation.
Morphology and the Technologist
When instrument-generated flags mandate slide review, it is imperative that instrument-attending technologists be well trained in morphologic hematology. Such technologists should be familiar with abnormal red shapes, including those induced artifactually. Technologists should be especially alert to fragmented forms indicating angiopathic hemolytic anemias--notably those associated with disseminated intravascular coagulation. The nuances of other poikilocytoses have been amply received (20, 21).
Attending technologists should also be alert to the presence of immature WBCs and blasts. Failure to recognize or detect blasts, and hence myelodysplastic syndromes or acute leukemias, in sentinel slides could lead to delayed diagnosis and therapy. Patients with untreated acute myelogenous leukemia have a projected survival time of ~3 months (22). Failure to recognize or detect the microgranular promyelocytes of variant acute promyelocytic leukemia may delay treatment in a disorder that requires rapid intervention. A procoagulant tissue factor associated with the granular fraction of leukemic promyelocytes promotes disseminated intravascular coagulation such that >90% of patients present with a hemorrhagic diathesis. Death is not infrequent before any treatment can be initiated (23).
To the extent that significant series of missed nonautomated, standard, morphologic diagnoses of leukemia have not been published, it is difficult to make comparisons with the diagnostic record of automated instruments. It is probably fair to state that most if not all modern instruments perform as well as the Coulter STKS Analyzer (1) in detecting virtually all cases of acute leukemia in blood, armed as they are with a plethora of suspect and definitive flags.
Future of Automated Hematology
Automated hematology 40 years later is still in its infancy. The recognition of left shifts, blasts, circulating lymphoma cells, or atypical lymphocytes remains an exercise in flagging, providing a mandate for sentinel slide review. Tacit acknowledgement of the current state of the art is provided by the Advia[TM] 120 Hematology System, which reserves a box on its scatter-plot grid for cells not recognized by cytochemistry. It properly labels them "large unstained cells (LUCs)". Undue activity in this box demands prompt morphologic review.
A pandemic problem among most analyzers performing differential white cell counts is an inability to distinguish between small lymphoid blasts, circulating small lymphoma cells, and normal lymphocytes. Thus, flags may not be displayed even in the presence of abnormal lymphoid populations. It has been suggested that automated laboratories set policies for reviewing sentinel slides in patients with unexpected absolute lymphocytosis to avoid missing malignancies in the lymphoid system (24).
Undoubtedly, in the first few decades of the new millennium advancing technology will routinely identify more white cell subtypes and even blast lineage while reducing the need for flags. Recognition of specific poikilocytes will also evolve, providing a major leap forward in the detection of rarer anemias or specific organ diseases. At the present time, automated hematology cannot provide these services.
I thank LoAnn Peterson, MD, for information on flagging practices at Northwestern Memorial Hospital, Chicago, IL; Geoffrey A. Witrak, MD, for flagging practices at St. Mary's/Duluth Clinic, Duluth, MN; and P. Joanne Cornbleet, MD, at Stanford University Medical Center, Stanford, CA for advice on technical matters pertaining to automated analyzers.
(1.) Hoyer JD, Fisher CP, Soppa VM, Lantis KL, Handon CA. Detection and classification of acute leukemia by the Coulter STKS hematology analyzer. Am J Clin Pathol 1996;106:352-8.
(2.) National Committee for Clinical Laboratory Standards. Reference leukocyte differential count (proportional) and evaluation of instrument methods; approved standard. NCCLS Document H2O-A. Villanova, PA: NCCLS, 1992;12:1-55.
(3.) Krause JR. The automated white blood cell differential: a current perspective. Diagn Hematol 1994;8:605-16.
(4.) Buttarello M, Gadotti M, Lorenz C, Toffalori E, Ceschini N, Valentini A, Rizzoti P. Evaluation of four automated hematology analyzers. Am J Clin Pathol 1992;97:345-52.
(5.) Burchert-Graeve, Kock R. Automated leukocyte differentials in 292 patients with leucopenia: an evaluation of the Abbott CELL-DYN 3500 (CD3500) haematology analyser. Clin Lab Haematol 1996; 18:253-9.
(6.) Petit MC, Fournier M, Boniface B, Ghavaert C, Caillaux M, Chretien MC, et al. The Abbott Cell-Dyn 3500 and the Coulter STKS: automated enumeration and differentiation of leukocytes in an HIV-positive population. Lab Hematol 1995;1:135-42.
(7.) Grimaldi E, Scopacasa F. Evaluation of the Abbott CELL-DYN 4000 hematology analyzer. Am J Clin Pathol 2000;113:497-505.
(8.) Strobel SL, Panke TW, Zelenski K. Hemoglobin C acquired via a blood transfusion: detection by automated blood counter. Arch Pathol Lab Med 1987;111:565-8.
(9.) Ree S, Williamson D, Garrett RW. Detection of abnormal hemoglobins by the Technicon H6000 automated cell counter. Br J Haematol 1985;59:734-6.
(10.) Booth R, Mead SV. Resistance to lysis of erythrocytes containing haemoglobin C detected in a differential white cell counting system. J Clin Pathol 1983;36:816-8.
(11.) Honda SAA, Bhagavan NV, Sugiyama CE, Gallaty CR, Dublin R, Flegal I, et al. Hemoglobinopathies detected by CBC analysis and HPLC hemoglobin A1~ analysis. Lab Med 1994;25:176-81.
(12.) Cornbleet J. Spurious results from automated hematology cell counters. Lab Med 1983;14:509-14.
(13.) Cornbleet PJ, Myrick D. Unusual RBC indices in diabetic patient. Lab Med 1983;14:444-5.
(14.) Ahmed P, Minnach V, Michael JM. Platelet satellitosis with spurious thrombocytopenia and neutropenia. Am J Clin Pathol 1969;69:473-4.
(15.) Rappaport ES, Helbert B, Beissner RS, Trowbridge A. Automated hematology: where we stand. South Med J 1988;81:365-70.
(16.) Jandl JH. Blood. Textbook of hematology, 2nd ed. New York: Little, Brown and Company, 1996:53pp.
(17.) Wallerstein R0. Laboratory evaluation of anemia. West J Med 1987;146:443-51.
(18.) Beutler E. The common anemias. JAMA 1988;259:2433-7.
(19.) Bergin JJ. Evaluation of anemia: getting the most out of the MCV, RDW, and other tests. Postgrad Med 1985;77:253-69.
(20.) Ward PCJ. Investigation of poikilocytic normochromic normocytic anemia. 1. Round and elongated forms. Postgrad Med 1979;65: 215-23.
(21.) Ward PCJ. Investigation of poikilocytic normochromic normocytic anemia. 2. Spiculated forms. Postgrad Med 1979;65:229-35.
(22.) Tivey H. The natural history of untreated acute leukemia. Ann N Y Acad Sci 1955;60:322-58.
(23.) Awisati G, Cate JW, Mandelli F. Acute promyelocytic leukaemia. Br J Haematol 1992;81:315-20.
(24.) Cornbleet PJ, Myruk D, Levy R, Judkins S. Cell Dyn 3000 and Coulter STKS automated differential counters [Letter]. Am J Clin Pathol 1994;100:464-5.
Patrick C.J. Ward
[1.] Nonstandard abbreviations: WBC, white blood cell; RBC, red blood cell; Hb, hemoglobin; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; RDW, red cell distribution width; IDA, iron deficiency anemia; and ACD, anemia of chronic disease.
Department of Pathology and Laboratory Medicine, University of Minnesota, Duluth School of Medicine, 10 University Dr., Duluth, MN 55812. Fax 218-726-7559; e-mail firstname.lastname@example.org.
Received April 5, 2000; accepted May 12, 2000.
Table 1. Methodologic permutations in a group of modern cell analyzers. Instrument Methodology features (comments) Coulter STKS Impedance, conductivity, laser technology Coulter GEN-S Impedance, flow cytometry, conductivity, laser light scatter (reticulocyte analysis) Sysmex SE-9000 Direct current (resistance to cell volume); radio-frequency (cell density, size) Cobas Argos 5-diff Impedance, light absorption (halogen light (Roche) source) Technicon H-3 (Bayer) Cytochemistry, flow technology (reticulocyte analysis) Advia 120 (Bayer) Cytochemistry, flow technology (reticulocyte analysis) Cell Dyn 4000 (Abbott) Impedance, laser light scatter Table 2. Factors known to cause spurious laboratory results in hematology analyzers. Index Spuriously increased Spuriously decreased RBCs Hb In vivo hemolysis, hyperlipidemia, Clotting hyperbilirubinemia, WBC >50 000/[mm.sup.3] MCV Cold agglutinins, hyperglycemia, Cryoglobulins WBC >50 000/[mm.sup.3] MCHC Hyperlipidemia, cold agglutinins WBC >50 000/ [mm.sup.3] RDW Status post transfusion WBCs Nucleated red cells, platelet Clotting clumps, unlysed red cells, Cryoglobulins Platelets WBC fragmentation, severe Satellitism, microcytosis, Cryoglobulins clumping
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Beckman Conference|
|Author:||Ward, Patrick C.J.|
|Date:||Aug 1, 2000|
|Previous Article:||Advances in Clinical Laboratory Hematology. The twenty-third Arnold O. Beckman Conference in Clinical Chemistry.|
|Next Article:||Flow cytometry: Principles and clinical applications in hematology.|
|Collective voices, community change imperative for Congressional Black Caucus.|
|Arnold O. Beckman: One Hundred Years of Excellence.|