Q&A column, 9/16

Create PDF

Editor: Frederick L. Kiechle, MD, PhD

Submit your pathology-related question for reply by appropriate medical consultants. CAP TODAY will make every effort to answer all relevant questions. However, those questions that are not of general interest may not receive a reply. For your question to be considered, you must include your name and address; this information will be omitted if your question is published in CAP TODAY.

Q. We know we can count fewer than 100 cells for a manual differential if there is a very low white cell count. But if the white cell count is very high, should we count more than 100 cells? Some references state that >30,000 WBC/µL require a 200 cell differential, others >50,000 WBC/µL, and many do not mention at all the need to increase above 100 cells counted.

A. The manual differential cell count standard operating procedure may vary from laboratory to laboratory and so, too, the specific denominator of total leukocytes counted. Likewise, this denominator may not necessarily be fixed for all white blood cell counts.

One of the most commonly cited references used to support laboratory SOPs in this regard is Henry’s,¹ which suggests that a differential count of 100 cells is “usually made” owing to “practical reasons.”¹

As required, the absence of at least 100 cells on a given smear would preclude a denominator of at least 100. However, when the WBC count is more substantial, a marked drop-off in precision will be noted if the denominator does not increase in step. This observation was noted in a classical paper by Rumke,² who also pragmatically suggests that the operator begin with a 100 cell differential with the option of expansion to 200 should circumstances require it (the identification of rare cells, for example). In the context of extreme WBC, however, the logic of a higher differential cell count relates to an understanding of the Poisson distribution (see, for example, Statistics in Medicine, page 133).³

Suppose, for example, that we want to be certain to detect monocytosis over a variety of WBC values (which, for our purposes, has a normal relative range of two to eight percent). The normal range for WBC values is approximately 4,000–11,000 leukocytes/µL.¹

Now assume that the WBC count is reduced to 500 cells/µL but that the monocyte is relatively increased, say to 25; this produces the Poisson cumulative probability distribution function seen in Fig. 1.

In this case, we can clearly see that a count of just over 100 cells will give a correct monocyte count most of the time.

Now assume that the WBC count is markedly increased to 30,000 cells/µL and that the monocyte count is only slightly above normal (say, 10 percent); this produces the Poisson cumulative probability distribution function seen in Fig. 2.

In this scenario, to be certain not to miss a subtle monocytosis, a count of close to 3,000 cells would be required.

In this latter scenario, therefore, even a 200 cell count would be woefully inaccurate. In such circumstances, as is the standard operating procedure in our laboratory, a 200 cell count is verified with a comment noting the statistical inaccuracy of the differential. Because of the wide variation of clinical scenarios, there is no clear consensus in the literature about the thresholds for increasing the number of cells that should be counted in a differential when higher cell numbers are present. An understanding of the statistical reasoning behind the need to evaluate higher numbers of cells can inform the development of an SOP to address how to handle specimens in which a manual differential is performed and cell counts are high.

It should be noted, however, that most modern analyzers avoid statistical inaccuracy by defaulting to a much greater denominator than would be tenable by manual differential. For smaller or lower-throughput laboratories lacking access to modern technologies or for when a manual count is required that would not be supported by an automated differential, a clearly prescribed SOP and a comment of a low precision at extremes (both high and low cell numbers) of WBC would address the imprecision inherent in the counts.

1. Vajpayee N, Graham S, Bem S. Basic examination of blood and bone marrow. In: McPherson RA, Pincus MR, eds. Henry’s Clinical Diagnosis and Management by Laboratory Methods. 22nd ed. Philadelphia: Elsevier Saunders; 2011:509–535.
2. Rumke CL. Variability of results of cell differentiation in blood smears [in Dutch]. Ned Tijdschr Geneeskd. 1958;102(51):2505–2508.
3. Riffenburgh RH. Statistics in Medicine. 3rd ed. Waltham, Mass.: Elsevier; 2012.

Etienne Mahe, MD, Pathologist, Division of Hematology and Transfusion Medicine, Department of Pathology and Laboratory Medicine, University of Calgary, Member, CAP Hematology/Clinical Microscopy Resource Committee

Q. We were using Bayer Clinitest reagent tablets to test for reducing substances other than glucose as a quick, inexpensive, and noninvasive screening test in an infant’s urine. Bayer discontinued producing the tablets and they are no longer available. What is the recommended method to replace this screening for reducing substances?

A. Testing urine for reducing substances as a newborn screen for disorders of carbohydrate metabolism was a standard laboratory practice for decades. The goal of this screen was primarily to detect galactosemia, as other disorders of carbohydrate metabolism are rare and generally more clinically benign.1 Since the advent of newborn screening programs in the mid-1960s, testing for galactosemia has been mandated by all 50 states. However, rather than detecting reducing substances in urine, the current recommended method uses a blood specimen and detects activity of the deficient enzyme in galactosemia (galactose-1-phosphate uridylyltransferase, or GALT). This method has improved sensitivity and specificity compared with the previously used urine test. The urine test can be falsely negative in affected newborns because galactose may not be present in the urine of affected patients at all times. It may also be falsely positive in patients with kidney injury and/or liver disease.2 For these reasons, screening for carbohydrate metabolism disorders by the Bayer Clinitest reagent tablets, or any other similar method, is no longer recommended. Urine tests for carbohydrates can be used in lieu of this test as well.