(Chest. 2003;123:1181-1187.)
© 2003
American College of Chest Physicians
Variations in Pleural Fluid WBC Count and Differential Counts With Different Sample Containers and Different Methods*
Barrett D. Conner, MD;
Y. C. Gary Lee, MD;
Paul Branca, MD, FCCP;
Jeffrey T. Rogers, RRT;
R. Michael Rodriguez, MD and
Richard W. Light, MD, FCCP
* From the Department of Pulmonary Medicine, Saint Thomas Hospital and Vanderbilt University, Nashville, TN.
Correspondence to: Richard W. Light, MD, FCCP, Director of Pulmonary Disease Program, St. Thomas Hospital, 4420 Harding Rd, Nashville, TN 37202; e-mail: rlight98{at}yahoo.com
 |
Abstract
|
|---|
Objective: To compare the results of pleural fluid analysis for WBC counts and differential cell counts as follows: (1) counting performed manually vs that performed by an automated cell counter; (2) cells collected in different types of specimen containers; and (3) cell counts performed at 4 and 24 h postthoracentesis.
Methods: Twenty-eight pleural fluid samples were each collected in five different containers (ie, ethylenediaminetetraacetic acid (EDTA)-treated glass, citrate-treated glass, heparinized glass, plain glass, and plain plastic tubes). The WBC counts and differential cell counts were obtained manually (on the EDTA tube) and with an automated counter on all tubes within 4 h of collection, and again after 24 h of refrigeration.
Results: There was a close correlation between the WBC counts obtained manually and those obtained with the automated counter from the pleural fluid samples collected in the EDTA tubes (r = 0.92). With the automated counter, the pleural fluid WBC counts were similar among the three tubes containing anticoagulants, but the counts obtained from the tubes without anticoagulants were significantly lower. The differential cell counts obtained manually and those obtained with the automated cell counter differed substantially. Although the percentage of lymphocytes was similar, the automated counter was inaccurate in differentiating neutrophils from large monocytes or mesothelial cells. The WBC counts obtained within 4 h of collection and 24 h after collection were virtually identical.
Conclusions: The WBC counts obtained manually and with the automated counter from pleural fluid samples in EDTA tubes correlated very closely. The pleural fluid WBC count was lower if the pleural fluids had been collected in tubes without an anticoagulant. Automated WBC counts from pleural fluid specimens were inaccurate, possibly due to difficulty in separating neutrophils from monocyte/mesothelial cells. Refrigerated storage for up to 24 h had no significant effect on the total WBC count or on the WBC count differential regardless of the tube utilized.
Key Words: cell counts • differential • pleural fluid • thoracentesis
|
Introduction
|
|---|
The measurements of pleural fluid WBC count and differential count are helpful in the diagnosis and management of patients with pleural effusion. In general, a WBC count of > 1,000 cells/µL suggests an exudate, while most transudates have WBC counts of < 1,000 cells/µL. Parapneumonic effusions usually have WBC counts of > 10,000 cells/µL, however, similar WBC counts are also seen in pleural effusions that are related to pancreatitis, pulmonary infarction, collagen vascular diseases, malignancy, and tuberculosis.1
2
The WBC count differential is one of the most informative tests performed on pleural fluid samples and can aid significantly in identifying the etiology of many effusions.2
A predominance of neutrophils in an exudative pleural fluid sample indicates acute inflammation of the pleural surface, as can be seen in pneumonia, pulmonary embolism, subphrenic abscess, early tuberculosis, and also pancreatitis. The discovery of an exudative effusion with > 50% small lymphocytes is suggestive of a malignant disease, tuberculous pleuritis, or a pleural effusion after coronary artery bypass graft surgery. If a large percentage of lymphocytes is found, the pleural fluid may be sent for flow cytometry to determine whether a clonal cell population is present, as is seen in lymphoma. Mesothelial cells line the pleural cavity, become dislodged, and are a normal part of pleural effusions. Their presence or absence is useful diagnostically as they are usually absent in tuberculous pleural effusions and complicated parapneumonic effusions. These conditions are associated with a marked inflammatory response that covers the pleura and prevents the shedding of mesothelial cells into the pleural fluid. Eosinophils in the pleural fluid are most often the result of either air or blood in the pleural space. Other unusual diagnoses to consider in the presence of pleural fluid eosinophilia include asbestos-related diseases, drug reactions, or parasitic diseases such as paragonimiasis, hydatid disease, amebiasis, or ascariasis.3
A recent review of the English medical literature, revealed no published data with regard to standards for pleural fluid collection and transport for the determination of the WBC counts and differential cell counts. Moreover, there are no data regarding changes in the measured WBC counts or differential cell counts in pleural fluid specimens over time. Currently, at our institution, pleural fluids for WBC count determination and differential cell counts are submitted in tubes that have been anticoagulated with ethylenediaminetetraacetic acid (EDTA), citrate, or heparin, and in plastic or glass tubes without anticoagulants.
At most hospitals, cell counts and differential counts of body fluid samples are performed manually. Manual cell counts are dependent on the availability of skilled laboratory personnel, are costly, and often are available only during certain hours, making delays in cell count determinations common. Clearly, an alternative to manual cell counting that is accurate, expeditious, easily obtained, and more cost-efficient is attractive.
The present study had the following three objectives: (1) to determine whether the pleural fluid WBC count and differential count, as determined by an automated cell counter, are comparable to the pleural fluid WBC count and differential count determined manually; (2) to compare the pleural fluid WBC count and differential count obtained using samples collected in tubes of varying composition and with varying anticoagulants (ie, EDTA-treated glass, heparinized glass, citrate-treated glass, plain glass, and plain plastic tubes); and (3) to compare the measured WBC counts and differential counts from the various tubes measured within 4 h of collection and at 24 h postcollection.
Our hypotheses were as follows: (1) accurate pleural fluid WBC counts and differential counts could be obtained using the automated cell counter; (2) tubes containing an anticoagulant would provide more accurate pleural fluid WBC counts and differential counts than tubes without anticoagulants, because the cells might clump if no anticoagulant was present; and (3) a 24-h delay in performing WBC counts and differential counts on pleural fluid samples would not significantly affect the result.
|
Materials and Methods
|
|---|
Between March 16, 1999, and April 14, 1999, we prospectively studied 28 pleural fluid samples from 26 patients who had undergone thoracentesis under ultrasound guidance in the Department of Radiology. The study was approved by the Institutional Review Board of Saint Thomas Hospital, and written informed consent was obtained from each patient. The pleural fluid was obtained using a 17-gauge needle with a plastic catheter, the placement of which was directed under ultrasound guidance. Portions of each pleural fluid sample were placed in EDTA-treated, citrate-treated, heparinized, and plain glass tubes, as well as in the plain plastic tubes found in commercially available thoracentesis kits (Baxter; Deerfield, IL). The samples were refrigerated at 4°C until the WBC counts and differential counts were obtained.
The pleural fluid WBC count and differential count from each of the tubes was determined within 4 h of the patient undergoing thoracentesis, and again was performed at 24 h postthoracentesis using an automated cell counter (ADVIA-120; Bayer; Tarrytown, NY). Manual WBC counts and differential counts also were performed using a Neubauer hemocytometer (American Optical; Buffalo, NY), and a Wright stained blood smear, respectively, on the samples collected in the EDTA-treated glass tubes within 4 h of collection, and they were used as reference values.
Statistical Analysis
All data were analyzed with a statistics software program (Sigma Stat, version 3.0; Jandel Scientific; San Rafael, CA). The manual and automated WBC counts were compared using a paired t test and linear regression. The WBC counts in the five different tubes were compared utilizing the one-way analysis of variance with repeated measures. Results with p < 0.05 were considered to be significant.
|
Results
|
|---|
Total WBC Count
The total pleural fluid WBC counts obtained with manual and automated counting methods on the EDTA-treated fluid samples were nearly identical (r = 0.92; p < 0.001) [Fig 1
]. When the total WBC counts of the pleural fluids collected in the five different tubes were determined by the automated counter, there were significant differences (Table 1
). The median cell counts obtained from the specimens in the plastic tubes and in the glass tubes were significantly (p < 0.05) lower than those obtained in the EDTA-treated tubes. The median cell count in the glass tubes was only 55% of the median cell count in the EDTA-treated tubes. Moreover, the reduction in the cell count from patient to patient was not consistent (Fig 2
), as some of the values approached 100% while others were < 25%. When the three tubes with anticoagulation were compared, the citrate-treated tubes and heparinized tubes had significantly lower median WBC counts than the EDTA-treated tubes, but their median values were higher than those of the glass tubes.
Manual vs Automated Pleural Fluid WBC Differential Counts
In contrast to the results from the total pleural fluid WBC counts, the differential cell counts obtained manually and with the automated method were not closely correlated when the samples from the EDTA-treated tubes were compared. For example, the correlation coefficient between the percentage of neutrophils determined manually and that determined with the automated machine was only 0.63, with a standard error of the estimate of 12.8% (Fig 3
, top). In addition, the correlation coefficient between the percentage of mononuclear/mesothelial cells determined by the two methods was only 0.19 with a standard error of the estimate of 12.5% (Fig 3
, middle). However, there was an excellent correlation between the percentage of lymphocytes determined by the two methods (r = 0.92) [Fig 3
, bottom].
View larger version (13K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3. Top: pleural fluid neutrophil counts of samples stored in EDTA-treated tubes using manual vs automated counting methods. Middle: pleural fluid monocyte/mesothelial cell counts of samples stored in EDTA-treated tubes using manual vs automated counting methods. Bottom: pleural fluid eosinophil counts of samples stored in EDTA-treated tubes using manual vs automated counting methods.
|
|
The pleural fluid differential cell counts determined with the automated counter from the five different tubes differed significantly (Table 2
). In general, the neutrophil percentage and the mononuclear/mesothelial percentage tended to be higher in the EDTA-treated tubes, while the lymphocyte percentage and the eosinophil percentage tended to be lower in the EDTA-treated tubes. In general, the correlation coefficients were higher among the counts obtained using samples in the five different tubes with the automated counter than among the counts obtained with the automated counter and manual counting (Table 3
). Nevertheless, the correlation coefficients obtained among the counts obtained using the five different tubes with the automated counter were frequently < 0.90.
View this table:
[in this window]
[in a new window]
|
Table 3. Pearson Product Moment Correlation for Automated Differential Cell Counts from the Various Cell Types and Tubes*
|
|
Effect of Storage on Pleural Fluid WBC Counts
When the automated cell counts were repeated at 24 h and compared to the initial automated WBC counts from similar tubes (eg, EDTA-treated or citrate-treated), there were virtually no differences (Table 4
).
View this table:
[in this window]
[in a new window]
|
Table 4. Pearson Correlation Values for Total WBC and Differential Count Cell Types Determined With the Automated Counter Initially and After 24 h of Storage at 4°C*
|
|
When the differential cell counts that had been obtained initially and at 24 h using the automated cell counter were compared, the correlations were also good, except for the counts with the glass tubes (Table 4)
. Therefore, a 24-h delay in determining pleural fluid cell counts is acceptable, unless one is using glass tubes that have not been treated with an anticoagulant.
|
Discussion
|
|---|
The present study shows that the pleural fluid WBC counts obtained with manual and automated counting methods on the EDTA-preserved fluid samples were nearly identical. However, the tube in which the fluid was collected influenced the WBC count determined with the automated counter. The median cell counts from samples collected in tubes without anticoagulant treatment (ie, glass and plastic tubes) were significantly lower than those in samples collected in EDTA-treated tubes, with the median cell counts in glass tubes being only 55% of those in the EDTA-treated tubes. The differential cell counts obtained with the automated counter when compared to those obtained manually were not sufficiently accurate for clinical use. There was no significant change in the WBC counts or the differential cell counts when these measurements were delayed for 24 h.
Since their introduction, automated cell counters have become indispensable instruments in the clinical laboratory for the analysis of blood samples. However, analyses of other body fluids are currently performed manually. Manual counting is a costly, time-consuming, labor-intensive laboratory procedure, which is dependent on the availability of skilled laboratory personnel. Moreover, the manual method of determining the WBC differential count has been shown to be imprecise and inaccurate in studies comparing results to a reference differential counting method.4
Sampling error is the major cause of the imprecise measures, due to the relatively small numbers of cells counted manually. In order to achieve an acceptable degree of precision, at least 1,000 cells should be counted.5
The modern automated cell counters are capable of performing > 100 complete hematologic cell counts in an hour, greatly diminishing the hours of labor utilized. Since the automated cell counter analyzes 1,000 to 20,000 leukocytes from each sample, the accuracy and the reproducibility of the results are superior to counts obtained manually when only 100 cells are counted.5
The new generation of cell analyzers utilizes variations of flow cytometry alone or in combination with other modalities. The cell analyzer used in the present study (ADVIA 120; Bayer) obtained the WBC count by passing a column of cells, after the RBCs underwent hemolysis, through an orifice where they were counted using two-angle laser light scatter signals. The differential cell count with the cell analyzer used in the present study was obtained using the Perox method. With this method, the cells are initially incubated with a peroxidase solution that stains the granules. Then the cell suspension is passed through the flow cell where the absorption (peroxidase staining) and the forward light-scatter (size) of each WBC are measured. Then each cell is identified as a neutrophil, lymphocyte, monocyte, eosinophil, basophil, or large unstained cell from its position on a cytogram on which the peroxidase staining is plotted on the x-axis and the volume is plotted the y-axis. For the present study, the monocytes and the large undifferentiated cells were combined for the monocyte/mesothelial cell population.
To date, automated cell counters have been utilized mostly for the analysis of blood. However, this technology is also utilized in many automated urine analyzers, which provide for the cell counting of urine sediment.6
7
The automated particle-flow cytometers are more accurate and use less labor than visual microscopy.8
There have been few reported studies using automated cell counters for cell counts and differential counts of other body fluids. Ziebig and coworkers9
compared counts obtained with a hematology analyzer (Cell-Dyn 3500; Abbott Laboratories; Abbott Park, IL) and a urine flow cytometer (UF-100; Sysmex Corporation; Long Grove, IL) with counts obtained manually. They reported that in most cases the WBC counts from cerebrospinal fluid (CSF) were similar, but in CSF samples with high lymphocyte counts, the hematology analyzer gave a falsely high WBC count while the flow cytometer gave a falsely low rate. Van Acker and coworkers10
reported similar results from the CSF samples using the urine flow cytometer (UF-100). However, Aune and Sandberg11
reported that the automated cell counter (model H*2; Bayer) could replace visual microscopy in the analysis of CSF. Salinas and colleagues12
compared manual and automated WBC cell counts in 100 EDTA-preserved synovial fluid samples. They reported that the correlation coefficient between the two different methods was 0.986. In addition, they found that on repeated measures of the same sample, the coefficient of variation was less for the automated counts than for the manual.
The present study demonstrates that the total WBC counts from pleural fluid samples determined manually and using the automated counter are nearly identical when the fluids were collected in EDTA-treated tubes. It is noteworthy, however, that the total WBC count determined with the automated counter was dependent on the tube in which it was collected. If the pleural fluid had undergone anticoagulation, there was no significant difference in the total WBC count. However, the total WBC counts were significantly lower if the fluid had been collected in plastic or glass tubes without receiving anticoagulation (Table 2)
. The counts were nearly identical in the tubes with and without the anticoagulant for some specimens, while they were much lower in the tubes without the anticoagulant for other specimens (Fig 2)
. We speculate that lower counts in the tubes without anticoagulants were due to cell clumping. For determining a pleural fluid WBC count, we recommend that the fluid be collected in a tube with an anticoagulant, as opposed to the plain plastic tubes that are found in most commercially available thoracentesis kits.
The present study demonstrates that the differential cell count obtained with the automated counter differed significantly from the differential cell count obtained manually from the specimens preserved with EDTA (Fig 3)
. The primary problem appeared to be in differentiating the neutrophils from the monocyte/mesothelial cells. The automated counter both overcounted and undercounted the percentage of monocyte/mesothelial cells. This problem could possibly be corrected if the assignment of cells on the cytogram were changed.
The tube in which the specimen was collected influenced the differential cell count obtained with the automated counter. There was a higher percentage of neutrophils and monocyte/mesothelial cells and a lower percentage of lymphocytes in the EDTA-preserved specimen than in the other specimens. This observation suggests that the anticoagulant in the tube affects either the cell size or the peroxidase staining. An alternative explanation would be that there was clumping of the cells. If the cells other than the lymphocytes clumped and several cells that were clumped together were counted as one cell, then the percentage of lymphocytes would increase.
In the present study, the WBC counts obtained initially and after a delay of 24 h were virtually identical. These findings are compatible with those reported12
when cell counters have been used to determine the WBC count using samples of blood and synovial fluids.
In conclusion, the present study documents that the pleural fluid WBC counts obtained manually and with an automated cell counter from specimens preserved with EDTA were virtually identical. When the WBC count of a pleural fluid is being determined, the fluid should be anticoagulated because the WBC count is lower in plain plastic or glass tubes, presumably due to cell clumping. The pleural fluid WBC count is not affected if the fluid is stored for 24 h. Differential cell counts from pleural fluid samples that were determined with the automated cell counter were not sufficiently accurate for clinical use.
|
Footnotes
|
|---|
Abbreviations: CSF = cerebrospinal fluid; EDTA = ethylenediaminetetraacetic acid
This research was supported by The Saint Thomas Foundation, Nashville, TN, and Bayer Corporation, Tarrytown, NY.
Received for publication July 11, 2002.
Accepted for publication September 24, 2002.
|
References
|
|---|
- Light, RW, Macgregor, MI, Luchsinger, PC, et al (1972) Pleural effusions: the diagnostic separation of transudates and exudates. Ann Intern Med 77,507-513[ISI][Medline]
- Light, RW, Erozan, YS, Ball, WC, Jr Cells in pleural fluid: their value in differential diagnosis. Arch Intern Med 1973;132,854-860[CrossRef][ISI][Medline]
- Light, R Pleural diseases 4th ed. 2001,413 Lippincott Williams & Wilkins. Philadelphia, PA:
- Dotson M, Koepke J. Description and manual method performance vs. reference method. Paper presented at: NCCLS Differential Count Symposium, November 26–28, 1984, Arlington, VA
- Krause, JR The automated white blood cell differential: a current perspective. Hematol Oncol Clin North Am 1994;8,605-616[ISI][Medline]
- Fenili, D, Pirovano, B The automation of sediment urinalysis using a new urine flow cytometer (UF-100). Clin Chem Lab Med 1998;36,909-917[CrossRef][ISI][Medline]
- Hannemann-Pohl, K, Kampf, S Automation of urine sediment examination: a comparison of the Sysmex UF-100 automated flow cytometer with routine manual diagnosis (microscopy, test strips, and bacterial culture). Clin Chem Lab Med 1999;37,753-764[CrossRef][ISI][Medline]
- Delanghe, J, Kouri, T, Huber, A, et al The role of automated urine particle flow cytometry in clinical practice. Clin Chim Acta 2000;301,1-18[CrossRef][ISI][Medline]
- Ziebig, R, Lun, A, Sinha, P Leukocyte counts in cerebrospinal fluid with the automated hematology analyzer CellDyn 3500 and the urine flow cytometer UF-100. Clin Chem 2000;46,242-247[Abstract/Free Full Text]
- Van Acker, J, Delanghe, J, Langlois, M, et al Automated flow cytometric analysis of cerebrospinal fluid. Clin Chem 2001;47,556-560[Abstract/Free Full Text]
- Aune, M, Sandberg, S Automated counting of white and red blood cells in the cerebrospinal fluid. Clin Lab Haematol 2000;22,203-210[Medline]
- Salinas, M, Rosas, J, Iborra, J, et al Comparison of manual and automated cell counts in EDTA preserved synovial fluids: storage has little influence on the results. Ann Rheum Dis 1997;56,622-626[Abstract/Free Full Text]
TOP
Abstract
Introduction
