College of American Pathologists
Printable Version

  Q & A





June 2012

Fredrick L. Kiechle, MD, PhD

Question Q. Can endometrial hyperplasia be diagnosed in secretory phase endometrium? Are there criteria for diagnosing endometrial hyperplasia in secretory/secretory phase endometrium?

A. The diagnosis of endometrial hyperplasia in secretory endometrium should be made with caution and after consultation, as it is an uncommon occurrence. A study presented in abstract form reported endometrial hyperplasia or carcinoma arising in a functional, secretory phase background at a frequency of less than two percent.1 Medical/hormonal therapies for endometrial hyperplasia (PDF, 298 KB) may induce secretory-like changes, confusing the histological picture; therefore, knowledge of the clinical history is essential. However, in a secretory background, markedly crowded glands with little stroma and proliferative-type epithelium should fulfill the criteria for endometrial hyperplasia (complex or complex atypical). This is an instance in which the endometrial intraepithelial neoplasia criteria may be particularly helpful, as the lesional glands will display a different cytology than that of the background endometrium.2


1. Simon RA, Hansen K, Xiong JJ, et al. PTEN status and frequency of endometrial carcinoma and its precursors arising in functional secretory endometrium; an immunohistochemical study of 29 cases. Mod Pathol. 2012;25(Suppl 2): 1248A.

2. Mutter GL, Zaino RJ, Baak JPA, et al. Benign endometrial hyperplasia sequence and endometrial intraepithelial neoplasia. Int J Gynecol Pathol. 2007; 26:103–114.

Rochelle Simon, MD
Assistant Professor of Pathology and Laboratory Medicine
Warren Alpert Medical School of Brown University
Women and Infants Hospital of Rhode Island

Member, CAP Surgical
Pathology Committee

Question Q. When performing body fluid cell counts on a hemocytometer, what are acceptable limits of agreement between the two sides? Our laboratory uses a percentage, which works for the higher cell counts; however, when a cell count is low, such as one RBC counted on one side and zero on the other, this is a 50 percent discrepancy. Any advice?

A. In general, most body fluids should have cell counts performed using an automated method rather than a hemocytometer. The automated method has improved precision since many more cells can be counted compared with a manual method.1,2 However, both methods are imprecise when the cell count is low. One must do a manual count when the total cell count is below the analytic range of the instrument, or if the specimen contains clots or is highly viscous. Each laboratory needs to establish the lower limit below which a manual count must be performed, and this limit cannot be below that set by the instrument manufacturer.2 In addition, it is recommended that only specimen types that have been cleared for use should be used, as detailed by the instrument manufacturer. If the laboratory wishes to perform testing on a non-cleared specimen type, then a more detailed validation study must be completed to ensure the accuracy and reliability of results (this would then be considered a laboratory- developed test).2

Hemocytometer counts are performed in duplicate to help improve the precision of the measurement. The counting procedure itself is designed to increase accuracy by increasing the number of squares counted as the count goes down. In addition, it is recommended that a control specimen or procedural control be used for each eight hours of patient testing3 to further ensure the accuracy of the method used. Many laboratories have a rule for determining whether the duplicate measurements are sufficiently close to be accepted for reporting. In our laboratory, the side with the lower count must be within 10 percent of the higher side value. If the specimen is clotted or partially clotted, or if clumps of cells or debris are present, then a statement should be made noting that the count may be less accurate in this setting.

When the counts are very low, such as in normal cerebrospinal fluid, the imprecision of the measurement obtained when using the hemocytometer is high. The following example is given by Gregory P. Smith, MD, and Carl R. Kjeldsberg, MD, in their chapter “Cerebrospinal, Synovial and Serous Body Fluids” in Henry’s 20th edition.4 This example concerns normal cerebrospinal fluid from an adult, which would have a white count of 0–5 cells/μL:

“…using 18 large squares (1 mm2 each) in a Fuchs-Rosenthal type chamber with a depth of 0.2 mm, a total volume of 3.6 μL (18 × 0.2 μL per square) is examined. With 5 cells/μL, a total of 18 cells is counted. The coefficient of variation (CV) defined as 100 divided by the square root of the number of cells counted, is 100/square root of 18, or 24%. Therefore, ± 2 CV is about ± 48%. A Neubauer hemocytometer with nine 1 mm2 squares with a depth of 0.1 mm has a CV of 45% (± 90% for 2 CV) with the same cell concentration.”

It is clear from this example that the CV decreases as the count increases. For example, the CV would be around 10 percent when the total number of cells counted was 100/μL (100/square root of 100), which corresponds to a total cell count of 111 cells in a Neubauer chamber or 28 in a Fuchs-Rosenthal chamber. Therefore, once the total number of cells counted is above these levels, the 10 percent rule can be used.

Below those levels, one should use judgment. It seems reasonable that if both counts are within the normal range of 0–5 cells/μL (which would correspond to counts between 0–5 in the Neubauer chamber, and 0–18 in the Fuchs-Rosenthal chamber), then they should be accepted as “matching” within the limits of the measurement, as long as there are no concerns about the specimen or how the chamber was prepared. If any uncertainty exists, then duplicate counts should be repeated.

This approach covers the vast majority of specimens. Those left are the cases where the count is elevated, but not to the point where the 10 percent rule can be used. Depending on preference, one could establish a sliding scale for different levels using a statistical approach, or, more practically, choose a simple rule of thumb, such as ±10 cells, which is what some laboratories use and should suffice clinically.


1. Galagan KA, Blomberg D, Cornbleet PJ, Glassy EF, eds. Color Atlas of Body Fluids. An Illustrated Field Guide Based on Proficiency Testing. Northfield, Ill.: College of American Pathologists; 2006.

2. Clinical and Laboratory Standards Institute. Body fluid analysis for cellular composition; approved guideline. CLSI document H56-A. Wayne, Pa.: CLSI. 2006.

3. College of American Pathologists, Commission on Laboratory Accreditation. Hematology and Coagulation Checklist. July 11, 2011 edition. Northfield, Ill.: College of American Pathologists. 2011.

4. Smith GP, Kjeldsberg CR. Cerebrospinal, Synovial and Serous Body Fluids. In: Henry JB, ed. Clinical Diagnosis and Management by Laboratory Methods. 20th ed. Philadelphia, Pa.: WB Saunders Co. 2001.

Katherine A. Galagan, MD
Director, Clinical Laboratories
Virginia Mason Medical Center Seattle

Question Q. Is there any protocol for reference ranges for trans-gender populations? We do not often encounter this problem because our physicians state either female or male. This time, though, the account ordered a prostate-specific antigen test and entered female. Our system is set to hold tests that normally would not be ordered. The doctor said this patient had a sex change and he wanted female on the report. We were able to “fix” this, but what about cases in which results are sex-dependent or a patient may be receiving hormone therapy?

A. This is a difficult question to answer because there are few studies on reference ranges among transgender patients. While it is well recognized that there are differences in reference ranges between males and females, it is not a simple matter of assigning normal ranges to the individual’s original gender (at birth) or new one (after surgery). Some laboratory tests, such as for male and female hormones, are clearly altered by surgery. Others are dependent on differences in muscle mass between men and women (for example, creatinine) and might not be altered. Given this lack of data, there is no guidance for the laboratory as to correct reference range to accompany a laboratory result on a transgender patient. Since gender is usually self-reported, the best strategy is to use the appropriate reference range for that listing. The laboratory should accommodate a request by the attending physician to change the gender listed in the patient’s demographic record because the attending is in the best position to evaluate the test result.

Alan H.B. Wu, PhD
Clinical Chemistry Laboratory
University of California
San Francisco

Dr. Kiechle is medical director of clinical pathology, Memorial Healthcare, Hollywood, Fla.