| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Division of Laboratory Medicine, Department of Pathology and Immunology, Washington University School of Medicine, Box 8118, 660 South Euclid Ave., St. Louis, MO 63110
aauthor for correspondence: fax 314-362-1461, e-mail eby{at}pathbox.wustl.edu
The accurate antenatal prediction of fetal lung maturity (FLM) based on results from amniotic fluid samples is of utmost importance in the prevention of neonatal respiratory distress syndrome and its complications. The current "gold standard" for the determination of FLM is the evaluation of phospholipids (i.e., measurement of lecithin/sphingomyelin ratio and quantification of phosphatidylglycerol) in amniotic fluid samples by thin-layer chromatography. These tests are, however, time-consuming and not continuously available at most institutions. Lamellar bodies are lamellated phospholipids that represent a storage form of surfactant (1). Because lamellar body diameter (range, 1–5 µm) is similar to that of small platelets, lamellar body counts (LBCs) can be obtained rapidly with use of the platelet channel of a hematology analyzer (2).
Recently, a consensus LBC protocol was published, and a FLM cutoff of 50 000/µL was suggested without discussion regarding the hematology analyzer used (3). The majority of published reports to date have used a Coulter brand of hematology analyzer to establish clinical decision limits (2)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13). The published experience with other hematology analyzers, such as the Sysmex NE-1500 (14), for obtaining LBCs is limited. One study used LBCs from three different analyzers (two from Coulter, one from Sysmex) to assess FLM without providing any evidence that the instrumentation was not a source of imprecision (9). Our objective was therefore to compare LBC concordance from the following four hematology analyzers: Coulter Gen-S (Beckman Coulter), Sysmex XE-2100 (Sysmex), Cell-dyn 3500 (Abbott Laboratories), and ADVIA 120 (Bayer Corporation).
Leftover amniotic fluid samples sent to the Barnes-Jewish Hospital Laboratory for physician-ordered FLM analysis during a 4-month period were analyzed. Amniotic fluid samples with visible blood and/or meconium contamination, measurable hemoglobin (>0.01 g/L; Coulter Gen-S), or insufficient volume (<500 µL) were excluded. Samples were stored at 4 °C. Forty-four amniotic fluid samples were evaluated. Uncentrifuged samples were assayed on the four hematology analyzers within 10 days of receipt (median, 4 days; range, 1–10 days). A previous study confirmed that results of lamellar body counts performed on amniotic fluid stored at 4 °C were stable up to 10 days (15). Human studies committee approval was obtained for this study.
Intra- and interassay imprecision (as CVs) was evaluated by two approaches. The intraassay imprecision for platelet counts was first determined with an EDTA-whole blood sample manipulated to obtain a platelet count of 
40 000/µL. Measured platelet counts were as follows: Coulter Gen-S, mean (SD), 39 000 (2400)/µL (n = 6; CV = 6.1%); Sysmex XE-2100, 43 200 (1300)/µL (n = 5; CV = 3.0%); ADVIA 120, 48 500 (1900)/µL (n = 6; CV = 3.9%); and Cell-dyn 3500, 36 500 (1600)/µL (n = 5; CV = 4.3%). Intraassay imprecision for LBCs was determined using a pooled amniotic fluid sample. Results were as follows: Coulter Gen-S, mean (SD), 18 400 (500)/µL (n = 7; CV = 2.9%); Sysmex XE-2100, 13 900 (600)/µL (n = 7; CV = 4.6%); ADVIA 120, 12 965 (855)/µL (n = 6; CV = 6.6%); and Cell-dyn 3500, 31 300 (2100)/µL (n = 7; CV = 6.7%). Interassay imprecision for the Coulter Gen-S was also determined with the manufacturer-supplied abnormal CBC control, diluted 1:1 with Isoton buffer (performed in duplicate on 10 separate days): mean (SD), 35 000 (1330)/µL (CV = 3.8%).
LBC results for the 44 different amniotic fluid samples are shown in Table 1
. Using published guidelines for FLM based on LBCs performed on Coulter hematology analyzers (3), we determined the concordance of LBCs between the three other hematology analyzers and the Coulter Gen-S (Table 1
and Fig. 1
, A–C). Fourteen were considered mature (Coulter GEN-S LBC 
50 000/µL), 21 intermediate (Coulter GEN-S LBC 15 000–50 000/µL), and 9 immature (Coulter GEN-S LBC 
15 000/µL). For the ADVIA 120, the LBC was expressed as platelet count plus red blood cell ghosts and fragments (calculated LBC), as recommended by the manufacturer. Of the three analyzers, the Sysmex XE-2100 showed the best concordance (86%) with the Coulter Gen-S (Fig. 1A
). The concordance of the ADVIA 120 (calculated LBC) was 78% (Fig. 1B
). When uncorrected ADVIA platelet counts were analyzed, the concordance with Coulter Gen-S was only 63% (data not shown). Finally, the concordance of the Cell-dyn 3500 with the Coulter Gen-S was 66% (Fig. 1C
).
|
|
To assess the difference between the Coulter Gen-S and the other three instruments, we derived Bland–Altman-like plots (Fig. 1
, D–F). The statistical difference was assessed using the least-squares regression method. With a slope of -0.32 (95% confidence interval, -0.39 to -0.26; P <0.001), the difference between the Sysmex XE-2100 and the Coulter Gen-S LBC values was negative (Fig. 1D
). For the ADVIA 120 (calculated LBC) values, the slope was -0.46 (95% confidence intervals, -0.60 to -0.32; P <0.001) compared with the Coulter Gen-S (Fig. 1E
). By contrast, the difference between the Cell-dyn 3500 and Coulter Gen-S LBC values was positive (slope, 0.76; 95% confidence interval, 0.66–0.87; P <0.001; Fig. 1F
).
The mechanisms underlying these differences are unknown but are likely related to the way lamellar bodies are counted on each instrument. The four hematology analyzers use different principles to identify platelets. The ADVIA 120 measures two light-scatter angles of particles as they pass through a laser beam (optical analysis). Platelets are identified based on their volume and refractive index (16). Platelet counts include platelets with volumes up to 60 fL and exclude other similarly sized particles, such as red blood cell fragments. Thus, the LBC count on the ADVIA 120 was derived as the sum of all platelet-sized particles measured in the PLT channel (calculated LBC). In contrast, the CBC channels on the Sysmex XE-2100 and the Coulter Gen-S use impedance technology. The Coulter method (conventional impedance) counts particles by detecting changes in electrical resistance when a particle in a conductive liquid goes through a small aperture (17). The size of the electrical pulse generated is proportional to the particle volume. Platelets are identified based on their volume (2–20 fL). The Sysmex technology is different from the Coulter method in that it simultaneously detects conventional (direct current) and radiofrequency impedance (18). The latter is thought to reflect intracellular changes. The Cell-dyn 3500 combines optical scatter and impedance to increase the accuracy of particle counting (19).
The hematology analyzers evaluated in this study produced accurate and precise platelet counts in reference populations (20) and thrombocytopenic patients (21) despite measuring different physical properties. However, our data indicate that the concordance among instruments for enumerating lamellar bodies is poor. For example, eight ADVIA 120 and five Sysmex XE-2100 LBCs indicated intermediate FLM when the Coulter Gen-S LBC indicated maturity (50 000/µL), and seven Cell-dyn 3500 LBCs indicated mature fetal lungs when the Coulter Gen-S LBC indicated intermediate FLM (15 000–50 000/µL). Clearly, applying the Coulter LBC cutoff values for FLM to other brands of hematology analyzers could change the positive or negative predictive value of the measurement (depending on instrumental bias), either of which may have adverse clinical consequences.
In summary, it is clear that different hematology analyzers count lamellar bodies differently. It will be necessary to establish analyzer-specific LBC clinical decision limits that are confirmed by outcome-based studies.
Acknowledgments
We thank Dr. Curt Parvin (Department of Pathology & Immunology, Washington University School of Medicine, St. Louis, MO) for help in statistical analysis and Bayer Corporation, Diagnostics Division (Tarrytown, NY), for providing an ADVIA 120 hematology analyzer for clinical research and for technical assistance.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
