Ask for a quotation at This email address is being protected from spambots. You need JavaScript enabled to view it.

M. E. Urusova, V. A. Golovachenko, E. P. Gitel, V. G. Kolodko, N. D. Fanchenko and D. G. Polintcev

EVALUATION OF IMMUNOASSAY KITS BY FOUR MANUFACTURERS (Alcor BIO INC., ROCHE, DPC AND BAYER) FOR DETERMINATION OF CERTAIN ANALYTE CONCENTRATIONS IN HUMAN SERA.

Laboratory of infection disease diagnostics, Laboratory of hormonal diagnostics, Alcor Bio Inc., Saint-Petersburg; Hormonal Research Department, Clinic of Obstetrics and Gynecology, I. M. Sechenov’s Moscow Medical Academy, Moscow; Laboratory of Endocrinology, Scientific Center of Obstetrics, Gynecology and Perinatology, RAMS, Moscow. 

Enzyme immunoassay technique [6] is widely used for clinical quantification of numerous antigens and antibodies in human sera. The main advantages of this method are rapidity and simplicity in use, high sensitivity and specificity, and absence of biohazards in kit contents. Currently world market has great assortment of kits for quantification of various analytes (hormones, tumor markers etc.) produced by different manufacturers.

The purpose of this study was to compare concentrations of some analytes measured by commercially available kits of four manufacturers: Alcor Bio Inc., Diagnostic Products Corporation (DPC), Roche and Bayer Corporation. We used kits for quantitative determination of cortisol, progesterone, testosterone, triiodothyronine, thyroxine, human chorionic gonadotropin (hCG), prolactin, alpha-fetoprotein (AFP), luteinizing hormone (LH), follicle-stimulating hormone (FSH) and thyroid-stimulating hormone (TSH). We used quality control sera Lyphochek^® Immunoassay Plus Control (Bio-Rad Laboratories, Hercules, California) for additional verification (see Table 1).

Materials and methods. Subjects and serum samples. All sera were collected in Hormonal Research Department, Clinic of Obstetrics and Gynecology, I. M. Sechenov’s Moscow Medical Academy (further abbreviated as HRD) and in Laboratory of Endocrinology, Scientific Center of Obstetrics, Gynecology and Perinatology, RAMS (further abbreviated as LE). All individuals underwent medical examinations in these hospitals. Preliminary certain analyte concentrations were determined using commercially available kits in HRD and LE (see Table 2). Then serum samples were delivered to our laboratory (Alcor Bio Inc.) for further investigations.

We used control sera Lyphochek^® Immunoassay Plus Control (Bio-Rad Laboratories, Hercules, California) of three levels (Cat. No. 370, Lot No. 40090) for quality control of the assays performed.

Methods for analyte quantification. Kits used in this study are listed in Table 3. All procedures were performed according to attached manuals [8-11].

EIA kits for automated analyzer Immulite^® use polystyrene beads as solid phase, alkaline phosphatase as enzyme, and chemiluminescent substrate. [10].

EIA kits for automated analyzer Roche Cobas Core use polystyrene beads as solid phase, horseradish peroxidase as enzyme, and 3,3¢ ,5,5¢ -tetramethylbenzidin (TMB) as chromogen [11].

Table 1. Different analyte concentrations in control sera Lyphochek® Immunoassay Plus Control

1 NB: data are cited from manuals attached to control sera Lyphochek^® Immunoassay Plus Control (Bio-Rad Laboratories, Hercules, California; Cat. No. 370, Lot No. 40090).
² The normal level of mentioned above analytes are equal for Immulite и Immulite2000.

Table 2. Assay systems used in current study

Table 3. Used in current study kits.

EIA kits for automated analyzer ACS: 180 use direct chemiluminescence technology, paramagnetic beads as solid phase, and acridinium ester as chemiluminescent label. [9].

The kits by Alcor Bio Inc. are intended for semiautomatic method. They use microtiter plates as solid phase, horseradish peroxidase as enzyme, and TMB as chromogen [8].

Statistical analysis. We used Pearson’s correlation coefficient (r) [2,3,5] to evaluate the relations between data obtained in our laboratory, in HRD, and in LE. Comparisons of patient sera analyte levels in the various assays were by linear regression analysis of individual data points using the least square method. Also we plotted linear regression curve on scattergrams to characterize these relations (correlation segment, slope of regression curve) [2,3,5].

Results and discussion. 1. Determination of cortisol concentration. Comparison of cortisol concentrations measured in patient sera in different laboratories is presented in Figure 1. Evidently there is good correlation between data from our laboratory and LE, but correlation between data from our laboratory and HRD is lower. Moreover characteristics of linear regression are markedly different (correlation segments differ several fold although slopes of regression curve have comparable values). The possible explanation of this fact is that subjective divergence might exist in assay operations between laboratories using kits and analyzers Immulite. We can affirm in a high correlation level that measurements of cortisol concentration in sera by Alcor Bio Inc. and Immulite kits don’t differ regarding data from LE apart. Determination of Bio-Rad control sera cortisol concentrations confirms these results objectivity (see Table 1).

2. Determination of triiodothyronine and thyroxine concentrations. Comparison of triiodothyronine and thyroxine concentrations measured in patient sera in different laboratories is presented in Figure 2. We had in use sera delivered only from LE tested by Roche Cobas Core kits. We have no serum samples with low or high triiodothyronine and thyroxine levels since patients with marked pathology of thyroid gland hardly present among those of Center. Hence the scattergrams looks like dense cluster of individual data points without any apparent correlation. Low correlation coefficients rule out linear regression analysis performing. Therefore the kits can be compared using some oblique evidences i.e. comparing resulting diagnoses. Thresholds of normal levels for Alcor Bio and Roche kits are drawn in Figure 2 as straight lines. There are only a few instances (2 sera or 0.6%) when normal Alcor Bio levels of triiodothyronine correspond to Roche hypothyroidism. Also there are no cases when normal Alcor Bio levels of triiodothyronine correspond to Roche hyperthyroidism. Some sera with normal Roche levels of triiodothyronine correspond to low Alcor Bio levels (23 sera or 7%) and high Alcor Bio levels (7 sera or 2%). Thus we report that triiodothyronine concentrations measured by Alcor Bio and Roche kits are in the same diagnoses except for 32 sera or 9% of samples.

   There is group of samples (35 sera or 11%) having normal Alcor Bio thyroxine concentrations corresponding to Roche hyperthyroidism. Otherwise we report that thyroxine concentrations measured by Alcor Bio and Roche kits are in the same diagnoses.

   Bio-Rad control sera values of triiodothyronine and thyroxine measured by Alcor Bio kits are in Roche Cobas Core range (see Table 1) and very close to the mean values. Exception is triiodothyronine concentration in 2 and 3 level controls (Alcor Bio triiodothyronine concentrations are 1,2–2-fold higher than Bio-Rad control concentrations certified for Roche Cobas Core).

   

   Thereby we had to analyze cohort of sera with low and high triiodothyronine and thyroxine levels for further research of kit reliability. Such analysis would allow evaluating precise relations between hormone concentrations measured by these kits

   

         3. Determination of progesterone and testosterone concentrations. Comparison of progesterone concentrations measured in patient sera by Alcor Bio and DPC (Immulite) kits is presented in Figure 3a, by Alcor Bio and Bayer [ACS: 180] in Figure 3b. The same testosterone concentration data are shown in Figure 4a, b. First, there is high correlation between Alcor Bio and Bayer data versus Alcor Bio and DPC data (especially in case of testosterone concentrations). Second, both slope of regression curve is closer to 1.0 and correlation segment is closer to 0 in Alcor Bio/Bayer regression analysis versus Alcor Bio/DPC. So progesterone and testosterone concentrations measured by Alcor Bio resemble Bayer data nearer than DPC data. Alcor Bio data have lower levels (both progesterone and testosterone concentrations) and less correlation (especially in the case of testosterone concentrations) comparing DPC measured data.

We suppose that possible variation in technique principles; calibration standard and control material matrixes of the kits used would cause such low correlation coefficients.

Testosterone concentrations measured in Bio-Rad control serum level 1 by Alcor Bio kits are lower than those certified for compare kits (see Table 1); ranges for other level sera concentrations are similar. Progesterone concentrations measured in Bio-Rad control serum level 1 by Alcor Bio kits are lower than those certified for compare kits, unlike level 2 and 3 progesterone concentrations which are higher than those certified for compare kits (see Table 1). These observations are inappropriate to small slope of regression curve between Alcor Bio and DPC (Bayer) data. Kit differences mentioned above can account for this fact. Perhaps contributions of the man-made control serum matrix effect (which should be unequally measured by different assays) play some role in this situation (see below for further discussion).

4.  Determination of human chorionic gonadotropin (hCG), prolactin and alpha-fetoprotein (AFP) concentrations. High correlation coefficient between prolactin concentrations measured by different kits allows plotting linear regression curve (see Figure 5). At the left-hand of scattergram concentrations measured by Alcor Bio tend to increase in comparison with Roche (Cobas Core) data and to decrease at the right-hand of scattergram (small slope of regression curve and high correlation segment are observed).

Bio-Rad control sera prolactin values measured by Alcor Bio kits are in the range certified for Roche Cobas Core (see Table 1) although mean values are slightly higher. All three level control sera concentrations are in the left-hand of scattergram, agreeing with clinical specimen measurement results.

There is higher correlation between hCG concentrations measured by Alcor Bio and Bayer versus Alcor Bio and DPC (see Figure 6). Higher degree of coincidence is observed also in the former case (slope of regression curve is closer to 1.0 and correlation segment is smaller). Bio-Rad control serum hCG values measured by Alcor Bio kits are higher and agree with range of hCG concentrations specified for compare kits (except for control serum level 1, see Table 1). This observation is disagreeing with clinical specimen measurement data. Perhaps the contributions of Bio-Rad man-made control serum matrix effect play some role in this situation also (see below for further discussion).

Comparison of AFP concentrations measured by different kits result in sure conclusion that analyte quantification reliability by the three manufacturer’s kits is equivalent (in couples Alcor Bio/Roche and Alcor Bio/DPC).

Bio-Rad control sera AFP values measured by Alcor Bio kits are in the range and very close to mean values certified for Roche Cobas Core (see Table 1). Together they are much higher than mean values and even not coincide to concentration range certified for Immulite2000. Furthermore Bio-Rad control sera AFP values measured by Roche Cobas Core and Immulite2000 similar among themselves (except for level 3). All this facts can be excellent illustration of our suggestion about interference of control sera matrix effect with measurement techniques.

The effect of different components in biological materials on analytical techniques is termed the matrix effect. Extremely complex and variable mixture of proteins, carbohydrates, lipids, small molecules and salts constituting the sample is source of the matrix effect [1,4,7]. Lyphochek^® Immunoassay Plus Control is prepared from human serum, with added constituents of human origin, pure chemicals and therapeutic drugs. In such “cocktail” substances not existing simultaneously in real serum can be mixed. Also 1^st level serum has minimal analyte concentrations and therefore should be free from number of components existing in normal human serum. Altogether Lyphochek^® Immunoassay Plus Control can be source of the matrix effect leading in differences between analyte recognition in control sera and subject of inquiry (human sera). A wide dispersion of analyte concentrations measured in the same control serum by different laboratories (and by different kits) can be one more oblique demonstration of our hypothesis [12].

These control sera is very useful for intralaboratory and interlaboratory quality control of the particular kit owing to high reproducibility of analyte concentrations measured by the same kit.

5. Determination of luteinizing hormone (LH), follicle-stimulating hormone (FSH) and thyroid-stimulating hormone (TSH) concentrations. LH, FSH and TSH concentrations measured in patient sera are shown in Figures 8-10 respectively. Good correlation and coincidence between all three kits were observed with high validity. We note that coincidence between LH concentration  Alcor Bio and Roche are slightly higher, and between FSH concentration measured by Alcor Bio and Bayer slightly higher, although consistency between the rest pairs is also fine enough. Determination of Bio-Rad control sera LH and FSH concentrations confirms these results objectivity (see Table 1). Bio-Rad 3^rd level control serum TSH concentrations measured by Alcor Bio kits are 2–3-fold lower than those certified for compare kits. This observation is oblique demonstration of our hypothesis about influence of man-made Bio-Rad control serum matrix effect on analyte quantification once more.

Conclusion. There is consistency and correlation between analyte concentrations measured by four commercial available kits from different manufacturers.

Coincidence degrees slightly decrease as follows: Alcor Bio/Bayer, Alcor Bio/Roche, and Alcor Bio/DPC.

Our observation about divergence between cortisol concentrations measured in two different laboratories by automated analyzer Immulite comparing to Alcor Bio data can trigger further development of the problem of work uniformity in different clinical laboratories. Comparison of the analyte concentrations measured in the same sera by different laboratories, investigation of such fluctuation reasons etc. will help us in this problem resolving.

No categorical conclusion can be deduced from results of this preliminary study. All possible variations associated with assay performing in different labs moreover in different cities (that is to say sample freezing/thawing, transportation etc.) should be eliminated to discuss about similarity or unlikeness of some kits. Therefore we are going to perform the analogous study within the single laboratory determining analyte concentration in the same sera by different kits simultaneously (without freezing in all study period).

All the same, analyte concentrations measured by four investigated kits are consistent each other with high validity. Thus all this kits can be very useful tools in clinical diagnostics practice with similar informativity and reliability levels.

LIST OF REFERENCE

1. Buttner J. // Scand. J. Clin. Lab. Invest. Suppl. – 1991. – N 205. – P. 11–20.

2. Glotov N.V., Zhivotovski L.A., Khovanov N.V. et al. Biometrics: Study guide. – Leningrad, 1982.

3. Lakin G.F. Biometrics. – Moscow, 1990.

4. Selby C. // Ann. Clin. Biochem. – 1991. – N 36. – P. 704–721.

5. Sergienko V.I., Bondareva I.B. Mathematical statistics in clinical observations. – Мoscow, 2000.

6. Theory and practice of immunoenzime assays / under the editorship of. A.M. Yegorov, A.P. Osipov, B.B. Dzantiev et al.– Moscow, 1991.

7. Wood W.G. // Scand. J. Clin. Lab. Invest. Suppl. – 1991. – N 205. – P. 105–112.

8. Work manuals to Alcor Bio kits.

9. Work manuals to kits for automatic analyzer ACS: 180^® (Automated Chemiluminescence System) © Bayer Corporation.

10. Work manuals to kits for automatic analyzer Immulite^®.

11. Work manuals to kits for automatic analyzer Roche Cobas^® Core.

12. Work manuals to Lyphochek^® Immunoassay Plus Control © Printed in USA.