| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Klinikum der Philipps–Universität Marburg, Department of Clinical Chemistry and Molecular Diagnostics, 35033 Marburg, Germany
aauthor for correspondence: fax 49-6421-2865594, e-mail hg.wahl{at}med.uni-marburg.de
In a recent Technical Brief on the IFCC-recommended method for lactate dehydrogenase (LD) measurement, Bakker et al. (1) questioned the reliability of this method for heparin samples. According to their findings, the IFCC method produced an excessive frequency of significantly discordant duplicate measurements (17.8%; 27 of 152 samples) compared with the formerly applied French Society (SFBC)-recommended method (1.4%; 2 of 140 samples). Neither increased centrifugation time and speed or the use of plasma separators to reduce the plasma contamination with blood cells, nor LD measurements without concurrent measurements of other analytes to eliminate possible interferences lowered the reported high rate of duplicate errors. The problem was also not attributable to the clinical chemistry analyzer because measurements on both Hitachi 911 and 717 analyzers (Roche Diagnostics GmbH) showed results similar to those obtained with the previously used Modular analyzer (Roche Diagnostics GmbH). Only when serum samples or secondary tubes for centrifuged plasma samples were used was there a significantly lower frequency of duplicate errors: 2.6% (3 of 114) for serum and 1.1% (1 of 94) for secondary plasma tubes. The authors pointed out that the differences in buffer composition of the IFCC (pH 9.40 ± 0.05) (2) and SFBC (pH 7.4–7.8) methods influenced the integrity of platelets and erythrocytes. In this case, the presence of cells might play a role in the high frequency of duplicate errors.
We have been measuring LD activity on three Roche Hitachi 917 analyzers according to the recommendations of the IFCC (2) in our laboratory since February 2003. The within-run imprecisions (CV) for a heparin-plasma sample (n = 20) were 0.8% (197.4 U/L) and 0.7% (321.3 U/L), and the between-run CV were 1.4% for the Roche Precipath U (178.3 U/L; n = 28) and 1.1% for the Roche Precinorm U calibrators (241.0 U/L; n = 28). During this period of time we observed no obviously erratic LD measurements for controls or patients when several samples from the same patient where analyzed within a few hours for follow-up. All measurements were routinely performed with plasma samples from plastic lithium-heparin-gel tubes (Monovette prod. no. 03.1631.001; Sarstedt). Blood sample collection was done without vacuum drawing. After centrifugation for 10 min at 3000g, the primary tubes were used for analysis. To reevaluate the reliability of the IFCC-recommended LD method in our laboratory with special regard to the reported high frequency of duplicate errors, we performed 140 LD duplicate measurements and studied the possible interferences from platelet contamination, cuvette and reagent probe/stirrer carryover effects, and different primary plasma tubes. To avoid long time delays between duplicate measurements, the analyses of the 140 samples were performed in the same run in groups of 5 with reruns.
The statistical analysis of the results from the duplicate measurements was calculated according to the Bland–Altman procedure (3)(4)(5). Fig. 1
shows the Bland–Altman plot with the means of the duplicate LD measurements (U/L) plotted against the absolute differences between duplicate measurements (U/L). For the differences the values of the second measurements were subtracted from the values of the first. The limits of agreement [mean (2 SD), 0.3 (26.6) U/L] (3)(4)(5) are shown as solid lines in Fig. 1
, and the mean of the differences as a dashed line. The frequency of duplicate errors [differences exceeding mean (2 SD)] under these conditions was 3.6% (5 of 140). This is more than the reported 1.4% for the SFBC method but less than the 17.8% reported for the IFCC method (1). The 95% confidence interval (-1.95 to 2.54 U/L; mean ± 2 SE) includes 0, so there is no evidence of systematic bias (3)(4)(5). [When comparing our data with the results published by Bakker et al. (1), please be aware that Bakker et al. introduced a conversion factor for the IFCC method.]
|
The interferences from carryover effects have been reported before (2)(6). To test cuvette carryover effects, we measured LD activity in patient samples in plastic lithium-heparin-gel tubes. The Hitachi analyzer we used was the Hitachi 917 Rack analyzer, which uses a cuvette ring of 160 cuvettes. Therefore, after cleaning and rinsing the cuvettes, we measured 160 LD samples to use every cuvette of the analyzer in one series of measurements. After that, 160 subsequent measurements of alanine aminotransferase (ALT), LD, aspartate aminotransferase (AST), and again LD were done. Thus, each cuvette was used for subsequent measurements of LD, ALT, LD, AST, and LD. Use of 160 samples in each run meant that each of the cuvettes had been used once for each analyte and that cuvette rinsing was performed only once in between. This was controlled by reporting both the cuvette numbers and the corresponding results. Interferences from previous AST or ALT measurements were expressed as absolute and relative differences (LDAST or ALT - LDclean)/LDclean for each measurement (Table 1
). The LD values measured after ALT or AST measurements tended to be higher than the first LD values with the "clean analyzer". However, the absolute differences were small, with means of 3.3 U/L for AST and 1.4 U/L for ALT (Table 1
). This small systematic difference (Table 1
) (3)(4)(5) may be the result of minimal carryover from LD-containing AST (LD 
3000 U/L; AST IFCC; Roche Diagnostics GmbH) and ALT (LD 3500 U/L; ALT IFCC; Roche Diagnostics GmbH) reagents (2).
|
Interferences from reagent probe/stirrer carryover effects were tested by measuring alternately LD and the suspected "offender" test (n = 100) without interruption from other tests. Possible offenders were defined as tests that could cause pH changes (lactate or calcium) or contamination with LD-containing reagents (ALT). All three tests caused small positive differences (Table 1
), with mean relative differences of 3.9% (lactate), 2.6% (ALT), and 1.7% (calcium). Rinsing the pipetter with an acidic solution (SMS/Acid wash; Roche GmbH) after the calcium test did not change the performance of the following LD measurements. These results point to minimal reagent probe/stirrer carryover effects, but would not explain the high frequency of duplicate errors reported by Bakker et al. (1), especially because the frequency was even higher when no other tests were run on the analyzer.
It has been shown that both platelet contamination of the samples and hemolysis can be responsible for LD measurements with poor precision in addition to higher LD activity values (7)(8). To investigate these effects, we performed LD measurements with serum-gel (Sarstedt Monovette prod. no. 02.1388), lithium-heparin-gel (same as given above), and lithium-heparin tubes (Sarstedt Monovette prod. no. 01.1604.400). Blood samples were collected from the same person in five different tubes and centrifuged differently to obtain various degrees of platelet contamination. Serum-gel, lithium-heparin-gel, and lithium-heparin samples were centrifuged for 10 min at 3000g or for 5 min at 500g. Platelet counts were done for all samples (Max M; Beckman-Coulter). IFCC-recommended serum samples had very low platelet contamination, even after reduced centrifugation (see Table 1
). Standard-centrifugation lithium-heparin samples contained platelet-poor plasma, whereas reduced centrifugation produced platelet-rich plasma (Table 1
). LD activity was measured (n = 77) for each of the five types of sample.
The higher platelet contamination of the samples in case of reduced centrifugation caused only a small increase of the within-run CV: 2.8% (142.9 U/L) and 3.2% (141.8 U/L) compared with 1.1–1.6% (133.9–165.2 U/L) for samples with low platelet contamination (Table 1
). The mean LD activity for standard-centrifugation lithium-heparin-gel samples was higher than the activity for the corresponding serum-gel samples (mean difference, 28.8 U/L) and the corresponding reduced-centrifugation lithium-heparin-gel samples (Table 1
). This may be the result of platelet destruction with subsequent release of intracellular LD. Lithium-heparin samples showed the same LD activity as lithium-heparin-gel samples despite higher platelet contamination. These results demonstrate that very high platelet contamination may have a small influence on the performance of the LD assay, but again, it would not account for the reported high frequency of duplicate errors.
Bakker et al. (1) found different frequencies of duplicate errors for heparin-plasma samples with (19%) and without separator (35%), and they could show that after heparin-plasma samples were transferred to secondary tubes (efficient mixing), the frequencies of duplicate errors dropped to 1.1%. We speculate that in the case of the Becton Dickinson heparin-plasma tubes used, both partial instability of the gel as well as inhomogeneities attributable to platelets and platelet aggregation might have caused the described problems. In combination with the specific sampling tubes, different variables, including blood sample collection, time between blood draw and centrifugation, reduced centrifugation, temperature, or specific analyzer features (e.g., rinsing program, sample and reagent volumes, or timing) could also contribute to the high frequency of duplicate errors.
We conclude that the IFFC method from Roche for LD measurement in heparin plasma is reliable, at least when performed under the conditions described here with Sarstedt Monovette tubes.
Acknowledgments
We thank B. Bruder-Burzlaff for excellent technical assistance.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  A. J. Bakker, A. Bakker, A. Bierma-Ram, J. T. Dijkstra, H. Renting-Wiering, H. Syperda, and A. Zijlstra Improved Reliability of Measurement of Lactate Dehydrogenase by IFCC Method in Heparin Plasma Clin. Chem., January 1, 2005; 51(1): 215 - 217. [Full Text] [PDF]
|
|
|
|
|
|
R. R. Miles, R. F. Roberts, A. R. Putnam, and W. L. Roberts Comparison of Serum and Heparinized Plasma Samples for Measurement of Chemistry Analytes Clin. Chem., September 1, 2004; 50(9): 1704 - 1706. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
