College of American Pathologists
Printable Version

  Q & A





cap today

April 2001

Q.  Our laboratory reports bacteria in the urine microscopic as rare, 1+, 2+, 3+, and 4+. As our hospital system includes many individual laboratories, I would like to standardize this evaluation by stating the actual number of bacteria/hpf that would be seen at each level. Can you provide references that give the actual number of bacteria/hpf that would be expected at each of these levels?

A.  Urine can be evaluated a number of ways for the presence of bacteria. The literature contains studies that evaluate uncentrifuged and centrifuged urine, both unstained and stained with Gram stain.

Since you indicate you are using a 1+ to 4+ system, it appears safe to assume you are referring to evaluation of centrifuged urine, as most studies using uncentrifuged urine simply evaluate for the presence of any bacteria per high power or oil immersion field. The most commonly used relation between the qualitative 1+ to 4+ systems and quantitative bacteria/hpf method in centrifuged urines is: 1+ = one to 10 bacteria/hpf; 2+ = 11 to 100 bacteria/hpf; 3+ = >100 bacteria/hpf; and 4+ = field packed with organisms.

The utility of either qualitative or quantitative reporting is its ability to predict the presence of a urine culture with >106 organisms, performed using standard methods. Several studies have reported the relation of the various bacteria/hpf levels in terms of how well it detects the presence of any positive urine culture (sensitivity) or how well it screens out urines that will not yield a positive culture (specificity). Using ≥1 bacteria/hpf (1+ or more) as a cutoff, sensitivity is 93 percent to 97 percent and the specificity 50 percent to 88 percent. Studies using the cutoff >10 bacteria/hpf (2+ or more) report sensitivity of 85 percent to 95 percent and specificity of 78 percent to 99 percent. Use of a cutoff of >100 bacteria/hpf (3+ or more) results in a sensitivity of 66 percent to 85 percent. Given the above data, using the following reporting system should provide clinicians with optimal information: one to 10 bacteria/hpf, 11 to 100 bacteria/hpf, and >100 bacteria/hpf.

It goes without saying that in addition to standardizing your reporting you need to standardize your sediment preparation and evaluation. Things that must be standardized are the volume analyzed, the rate of centrifugation, the duration of centrifugation, the volume of urine used to resuspend the sedimented pellet, the depth of the chamber used for evaluation, and the magnification used for examination.

1. Jenkins RD, Fenn JP, Matsen JM. Review of urine microscopy for bacteriuria. JAMA. 1986;255:3397-3403.

  Robert Novak, MD
Department of Pathology
Children's Hospital
Medical Center of Akron
Akron, Ohio

Microscopy Resource Committee

Q.  Can EDTA blood specimens be vortexed to obtain a platelet count when platelet aggregates are found?

A.  EDTA-induced platelet clumping is possibly the most common cause of pseudothrombocytopenia in the clinical hematology laboratory and can result in considerable difficulty in obtaining an accurate platelet count.1-6 The reported incidence ranges from one in 50 to one in 1,000 hospitalized patients, with a considerably lower incidence in healthy outpatients. This platelet clumping is an in vitro phenomenon caused by preexisting antibodies, presumably reacting with EDTA-modified antigens.4,7,8 Strictly speaking, it represents agglutination rather than aggregation, as it is not prevented by inhibitors of the platelet release reaction.9 In addition to pseudo- thrombocytopenia, platelet agglutination may cause pseudoleukocytosis due to the counting of platelet c clumps as leukocytes by automated analyzers.3

The most commonly employed solution to this problem is to redraw the specimen into a different anticoagulant, usually either sodium citrate or acid citrate dextrose; this resolves the problem in most cases.3 Obviously, however, this is inconvenient for the ordering physician and unpleasant for the patient. When a redraw into an alternative anticoagulant fails, another approach is to warm the specimen, as occasional cold platelet agglutinins are encountered.10 Finally, in particularly refractory cases, we have found that a freshly drawn specimen run immediately is helpful, suggesting that this may be a time-dependent phenomenon.

Whether vortexing is a viable alternative to redrawing the specimen has been investigated in a single study in the literature.5 In this study, 188 specimens with platelet clumps were vortexed for one to two minutes at or near the highest setting on the vortex. Either an aliquot or the entire tube could be used; the process of vortexing did not appear to adversely affect other CBC parameters. The specimens were reanalyzed on the automated analyzer within several minutes, and a new smear was made. These authors reported complete microscopic resolution of platelet clumping in 43.6 percent of samples, with an average 73 percent increase in the platelet count. An additional 49.5 percent of samples showed a partial resolution of clumping, with an average 111 percent increase in platelet counts. Finally, this maneuver was completely unsuccessful in 6.9 percent of samples. We have informally evaluated this procedure in our laboratory and have found it to be successful in about 50 percent of cases.

Based on these data, it seems reasonable to attempt to vortex samples as a first-line attempt to resolve platelet clumping. I would suggest that a post-vortex smear be made in each case to evaluate the extent of resolution of clumping. If no clumps are seen following vortexing and the platelet count has increased, I would suggest that the count may be reported. However, I would exercise caution in the situation when only partial resolution of clumping is observed, even if the platelet count increases substantially. My recommendation in this instance depends on the post-vortexing platelet count. If it is normal, I would suggest a comment that platelet clumping is present but that the platelet count is adequate. If the count remains decreased, a new specimen should be drawn in a citrate-containing anticoagulant. Finally, it is important to distinguish platelet clumps from microscopic fibrin clots, which may be seen to contain pale blue, amorphous, or fibrillary material in addition to platelets. Microscopic fibrin clots have been anecdotally observed to result in pseudothrombocytopenia in some cases; such specimens must be redrawn, although it is unnecessary to use an alternative anticoagulant.

1. Shreiner DP, Bell WR. Pseudothrombocytopenia: manifestation of a new type of platelet agglutinin. Blood. 1973;42: 541-549.
2. Payne BA, Pierre RV. Pseudothrombocytopenia: a laboratory artifact with potentially serious consequences. Mayo Clin Proc. 1984;59:123-125.
3. Lombarts AJ, de Kieviet W. Recognition and prevention of pseudothrombocytopenia and concomitant pseudoleukocytosis. Am J Clin Pathol. 1988; 89:634-639.
4. Manthorpe R, Kofod B, Wiik A, et al. Pseudothrombocytopenia. In vitro studies on the underlying mechanism. Scandinavian Journal of Haematology. 1981; 26:385-392.
5. Gulati GL, Asselta A, Chen C. Using a vortex to disaggregate platelet clumps. Laboratory Medicine. 1997;28:665-667.
6. Berkman N, Michaeli Y, Or R, et al. EDTA-dependent pseudothrombocytopenia: a clinical study of 18 patients and a review of the literature [see comments]. Am J Hematol. 1991;36:195-201.
7. Pegels JG, Bruynes EC, Engelfriet CP, et al. Pseudothrombocytopenia: an immunologic study on platelet antibodies dependent on ethylene diamine tetra-acetate. Blood. 1982;59:157-161.
8. von dem Borne AE, van der Lelie H, Vos JJ, et al. Antibodies against cryptantigens of platelets. Characterization and significance for the serologist. Curr Stud Hematol Blood Transfus. 1986;52:33-46.
9. Onder O, Weinstein A, Hoyer LW. Pseudothrombocytopenia caused by platelet agglutinins that are reactive in blood anticoagulated with chelating agents. Blood. 1980;56:177-182.
10. Watkins SP, Jr., Shulman NR. Platelet cold agglutinins. Blood. 1970;36:153-158.

  Steven H. Kroft, MD
Department of Pathology
The University of Texas
Southwestern Medical Center at Dallas

Microscopy Resource Committee

Miller Disk Clarification
The following paragraph from a previous "Q&A" response (December 2000, page 62) regarding the proper use of the Miller disk generated confusion:

To achieve acceptable levels of precision for absolute reticulocyte counts, large numbers of red cells must be counted, usually 1,000 to 2,000. This is a tedious and labor-intensive process. The Miller disk is designed to lessen this labor by reducing the number of cells that need to be counted. In essence, the Miller disk allows laboratory professionals to estimate the number of red cells in a microscopic field while directly counting the number of reticulocytes. When used properly, the Miller disk can increase the precision of the analysis while decreasing the time and effort required.

Specifically, readers inquired whether it was acceptable for fewer than 1,000 cells to be counted if a Miller disk was used. The following clarification is offered. The purpose of counting at least 1,000 red cells when manually enumerating reticulocytes is to achieve a relatively acceptable level of precision for the assay, based on the Poisson distribution. When using a Miller disk, it is important that reticulocytes be counted in a field that includes at least 1,000 red cells. This does not mean that 1,000 red cells need to be counted in the small square. In fact, a minimum of 112 cells must be counted in the small square, as this is equivalent to 1,008 red cells in the large square. However, it is imperative that the reticulocytes themselves be counted in the large square that contains at least 1,000 red cells. This recommendation is compatible with the CAP Hematology checklist question regarding the manual reticulocyte count, which states that the count must be "based on at least 1,000 red cells."

It should be noted that automated or semiautomated reticulocyte counting is available on many current hematology analyzers. These techniques provide far better precision for reticulocyte enumeration than manual methods and should be employed when possible.