Analytical performance of specific-protein assays on the Abbott Aeroset system.
Patient samples that had been sent to the laboratory for routine clinical chemistry investigations were collected and frozen over a 3-month period. The study was approved by the local Medical Ethics Committee. Samples were also obtained from the United Kingdom National External Quality Assessment Scheme and the Welsh External Quality Assessment Scheme. Assays were performed according to the manufacturer's recommendations on the Abbott Aeroset analyzer. Sample volume requirements were 2.0-6.5 [micro]L with reagent volumes <200 [micro]L. The Beckman Array Nephelometer and specific-protein reagents were used for comparison studies.
Precision was evaluated according to NCCLS protocol EP5-T2 (1) by use of single lots of reagents, calibrators, and controls (normal and high Biorad Immunochemistry controls, LVT1 and LVT2) for the entire study. Within-run imprecision (CV; n = 20) ranged from 0.4% for IgM (mean = 1.17 g/L) to 2.0% for apo A1 (mean, 1.32 g/L). Between-day and total imprecision were measured on two daily analytic runs for 20 days over a 67-day period by use of two control replicates per test (n = 80). The between-day CVs for control samples were <3.0% for all assays with the exception of apo B [LVT2, 6.5% (mean, 0.55 g/L); LVT1, 4.0% (mean, 1.09 g/L)]. The mean total CV across all specific-protein applications was 2.7%. The total CV for 19 of 20 controls was [les than or equal to] 5%, which met the minimum precision criteria on the basis of medical need (2,3). The exception was apo B (6.9% CV; mean, 0.55 g/L; minimum precision goal of 5.2%). Of the 10 specific-protein applications evaluated, all met the manufacturer's claims for total SD.
Linearity, evaluated by NCCLS protocol EP6-P (4), showed all assays linear across the ranges of their calibration curves: IgG, 1.25-37.36 g/L; IgA, 0.29-7.23 g/L; IgM, 0.14-3.43 g/L; transferrin, 0.54-5.38 g/L; haptoglobin, 0.11-2.77 g/L; prealbumin, 0.12-0.60 g/L; apo A1, 0.33-3.32 g/L; apo B, 0.26-2.56 g/L; complement C3, 0.14-3.22 g/L; and complement C4, 0.03-0.60 g/L.
Detection limits were determined by measuring 20 replicates of saline and 20 replicates of the lowest concentration calibrator and calculating the mean concentration of saline +2 SD of the lowest concentration calibrator. The limit of detection for assays was 0-0.02 g/L, with the exception of IgG at 0.11 g/L.
A within-run precision profile was constructed for IgA by analysis of 20 replicates of 15 patient samples with concentrations near the expected limit of detection (range, 0.01-0.11 g/L). Imprecision (CV) was then calculated for each patient sample and plotted as a function of observed analyte concentration. A 20% CV was seen at 0.02 g/L and a 10% CV at 0.03 g/L. The quoted lowest reportable concentration for the Beckman array IgA is 0.07 g/L.
Method comparison was performed according to NCCLS protocol EP9-T (5). Patient samples were measured on both the Aeroset and Beckman Array within a 2-h time span. A minimum of 52 serum samples were selected for each assay, with analyte concentrations evenly distributed over the assay range. Results of Bland-Altman analysis (6) are shown in Table 1. Overall, we found the Abbott Aeroset turbidimetric methods were in agreement with the Beckman Array nephelometric methods. The on-board Aeroset assay protocol for IgA and IgM includes an automated dilution step, so that each sample is analyzed undiluted and diluted, and a ratio is calculated. Results outside a user-defined acceptability criterion for the ratio (e.g., outside 0.5-1.5) suggests antigen excess, and patient results are flagged. Flagged samples are subsequently further diluted (1:10) manually with saline and reassayed. All antigen-excess samples observed for IgA and IgM were appropriately detected by the analyzer. The quoted upper limit for IgG before antigen excess is reached is 95 g/L. The highest concentration of IgG in the current study was 81.4 g/L.
Four replicates of each sample obtained from United Kingdom National External Quality Assessment Scheme and Welsh External Quality Assessment Scheme were measured on both the Abbott Aeroset and Beckman Array within a 2-h time span. The mean bias from the overall mean method value for each sample was calculated. Results for Aeroset were within -13.2% (IgM) to +7.4% (apo B) of overall method mean values and for Beckman Array, -0.65% (apo B) to +23.8% (apo A1). The bias observed in some external quality-assessment samples may have been attributable to matrix effects on stored samples because fresher nonfrozen patient samples showed better agreement between both instruments. A carryover study that was performed according to ECCLS guidelines (7), with the IgG assay as a representative specific protein, showed carryover <0.1% for a specimen containing 63.8 g/L IgG.
The effect of endogenous interference by either hemolysis or lipemia was evaluated according to NCCLS protocol EP7-P (8) on haptoglobin and apo B. Negligible interference (<10%) occurred on both assays with triglyceride (20% Intralipid) concentrations up to 10 g/L. Increasing the concentration of hemoglobin up to 10 g/L produced <10% interference on apo B. At a haptoglobin concentration of 0.96 g/L, the Abbott Aeroset haptoglobin showed less interference (-20%) at a concentration of 2.5 g/L hemoglobin than the Beckman Array, which showed interference (-40%), an effect that has been demonstrated previously on the Beckman Array (9).
In conclusion, the precision of 9 of the 10 Abbott Aeroset specific-protein methods studied in this evaluation met goals based on medical needs, and results were comparable to established methods on the Beckman Array. Consolidation of these specific-protein assays on a single chemistry analyzer can improve laboratory efficiency through decreased analyzer maintenance and decreased sample splitting. The small sample size required for these assays facilitates testing of pediatric samples.
We thank Abbott Laboratories for providing reagents, statistical support, and financial support for this work.
(1.) National Committee for Clinical Laboratory Standards. Evaluation of precision performance of clinical chemistry devices; tentative guideline EP5-T2. Wayne, PA: NCCLS, 1992:1-46.
(2.) Ricos C, Alvarez V, Cava F, Garcia-Lario JV, Hernandez A, Jimenez CV, et al. Current databases on biological variation: pros, cons and progress. Scand J Clin Lab Invest 1999;59:491-500.
(3.) Fraser CG. General strategies to set quality specifications for reliability performance characteristics. Scand J Clin Lab Invest 1999;59:487-90.
(4.) National Committee for Clinical Laboratory Standards. Evaluation of the linearity of quantitative analytical methods; proposed guideline EP6-P. Wayne, PA: NCCLS, 1986:509-72.
(5.) National Committee for Clinical Laboratory Standards. Method comparison and bias estimation using patient samples; tentative guideline EP9-A. Wayne, PA: NCCLS, 1995:1-36.
(6.) Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307-10.
(7.) European Committee for Clinical Laboratory Standards. Guidelines for the evaluation of analysers in clinical chemistry. Berlin: Beulth, 1986:17.
(8.) National Committee for Clinical Laboratory Standards. Interference testing in clinical chemistry; proposed guideline EP7-P. Wayne, PA: NCCLS, 1986:1-371.
(9.) Bossuyt X, Blanckaert N. Evaluation of interferences in rate and fixed-time nephelometric assays of specific serum proteins. Clin Chem 1999;45:62-7.
Ellen B. Duly *, George Barnes, Sandra Grimason, and Thomas R. Trinick
(Clinical Chemistry Laboratory, Ulster Hospital, Dundonald, BT16 1RH, Northern Ireland; * author for correspondence: fax 44-0-2890487131, e-mail firstname.lastname@example.org)
Table 1. Bland-Altman analysis of method differences with mean of methods. Lower 95% Upper 95% Bias, limit of limit of Mean range, g/L agreement agreement Assay n g/L (SD) (SD) (SD) 19G 54 1.22-78.29 0.85 -4.36 6.07 (0.37) (0.61) (0.61) 19A 52 0.05-45.4 -0.23 -4.32 3.87 (0.29) (0.49) (0.49) Igm 90 0.05-54.04 -1.86 -8.27 4.56 (0.34) (0.58) (0.58) Transferrin 60 0.82-3.65 -0.11 -0.27 0.05 (0.01) (0.02) (0.02) Haptoglobin 60 0.04-3.67 0.01 -0.25 0.26 (0.02) (0.03) (0.03) Prealbumin 60 0.10-0.55 -0.01 -0.03 0.01 (0) (0) (0) Apo A1 60 0.59-2.08 0.03 -0.03 0.08 (0) (0.01) (0.01) Apo B 60 0.38-2.88 -0.04 -0.10 0.03 (0) (0.01) (0.01) C3 60 0.72-1.93 -0.01 -0.07 0.05 (0) (0.01) (0.01) C4 60 0.14-0.44 0 -0.04 0.03 (0) (0) (0)
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Technical Briefs|
|Author:||Duly, Ellen B.; Barnes, George; Grimason, Sandra; Trinick, Thomas R.|
|Date:||Sep 1, 2001|
|Previous Article:||Serum C-reactive protein in Canadian Inuit and its association with genetic variation on chromosome 1q21.|
|Next Article:||Lead poisoning: a new biochemical perspective on the differentiation between acquired and hereditary neuroporphyria.|