Morrow J, Frei B, Longmire A, Gaziano J, Lynch S, Shyr Y, et al. Increase in circulating products of lipid peroxidation (F2-isoprostanes) in smokers. N Engl J Med ...
578
Technical Briefs
This work was supported by a grant from the American Diabetes Association and NIH grant M01-RR-00633. We thank R. Levine for useful discussions, Alicia Summers for technical assistance, and Beverly Huet Adams for statistical expertise. References 1. Stadtman ER, Oliver CN. Metal-catalyzed oxidation of protein. Physiological consequences. J Biol Chem 1991;266:2005– 8. 2. Levine RL, Garland D, Olivier C, Amici AA, Climent I, Lenz AG, et al. Determination of carbonyl content in oxidatively modified proteins. Methods Enzymol 1990;186:464 – 8. 3. Buss H, Can T, Sluis K, Domigan N, Winterbourn C. Protein carbonyl measurement by a sensitive ELISA method. Free Radic Biol Med 1997;23: 361– 6. 4. Morrow J, Frei B, Longmire A, Gaziano J, Lynch S, Shyr Y, et al. Increase in circulating products of lipid peroxidation (F2-isoprostanes) in smokers. N Engl J Med 1995;332:1198 –203. 5. Agil A, Fuller C, Jialal I. Susceptibility of plasma to ferrous iron/hydrogen peroxide-mediated oxidation. Demonstration of a possible Fenton reaction. Clin Chem 1995;41:220 –5. 6. Marangon K, Herbeth B, Artur Y, Siest G, Esterbauer H. LDL and VLDL composition and resistance to copper-induced oxidation are not modified in smokers. Clin Chim Acta 1997;265:1–12. 7. Reznick AZ, Cross CE, Hu ML, Suzuki YJ, Khwaja S, Safadi A, et al. Modification of plasma proteins by cigarette smoke as measured by protein carbonyl formation. Biochem J 1992;286:607–11. 8. Eiserich JP, Van der Vliet A, Handelman GJ, Halliwell B, Cross CE. Dietary antioxidants and cigarette smoke-induced biomolecular damage: a complex interaction. Am J Clin Nutr 1995;62:1490S–500S. 9. Marangon K, Herbeth B, Lecomte E, Paul-Dauphin A, Grolier P, Chancerelle Y, et al. Diet, antioxidant status and smoking habits in French male adults. Am J Clin Nutr 1998;67:231–9. 10. Yan LJ, Lodge JK, Traber MG, Packer L. Spectrophotometric method for determination of carbonyls in oxidatively modified apolipoprotein B of human low density lipoproteins. Anal Biochem 1995;228:349 –51.
Fig. 1. Time course of carbonyl (A) and lipid peroxide (B) formation during LDL oxidation. LDL, 0.5 g/L protein; Cu, 12.5 mmol/L. Mean of three different donors.
We also measured LDL carbonyl content, using this ELISA. During LDL oxidation, carbonyl increased continuously with time, paralleled lipid peroxides, as shown in Fig. 1, and was significantly correlated with lipid peroxides ( r 5 0.86; P ,0.001). Our results obtained with a copper/LDL protein ratio of 25 mmol/g protein are consistent with the results assessed spectrophotometrically with 20 mmol copper/g protein shown by Yan et al. (10 ), who reported a significant linear relationship between apo B LDL carbonyl content and its peroxidation as measured by thiobarbituric acid-reacting substances. Unlike their method, our method does not require a blank, precipitation, or delipidation. Both native and oxidized LDL can be assessed directly by immediate derivatization. Thus, the ELISA method is not only able to measure oxidative stress in circulating plasma, but can also be used to monitor LDL oxidation. In conclusion, because the ELISA requires a small sample volume and allows a large batch of samples to be run simultaneously, it may be useful in clinical studies aimed at comparing plasma as well as LDL oxidizability in different populations and also following antioxidant supplementation.
Technical Evaluation of Thyroid Assays on the Vitros ECi, Sharon Saw,* Sunil Sethi, and Tar-Choon Aw (Department of Laboratory Medicine, National University Hospital, Lower Kent Ridge Road, Singapore 119074; * author for correspondence: fax 65-7771757) The Vitros ECiTM analyzer (Ortho Clinical Diagnostics, Rochester, NY) uses a new enhanced chemiluminescence technology. The assay reagents and wells are supplied together in combined packs with calibration information stored on magnetic calibration cards with bar-coded calibrators. We evaluated the following thyroid assays: thyrotropin (TSH), free thyroxine (fT4), free triiodothyronine (fT3), thyroxine (T4), and triiodothyronine (T3). Noteworthy is that the novel free thyroid hormone assay uses an alternative methodology to the frequently used one-step analog or two-step methodology. The conventional analog system uses a tracer-labeled T4 conjugate. The fT4 assay uses a peroxidase-labeled antibody to T4. We assessed 872 consecutive subjects from a healthy adult population for TSH and fT4, and smaller cohorts of 120 and 146 subjects for the total hormone and fT3 reference intervals, respectively. The subjects conformed to accepted criteria for establishment of a reference population (1 ). In addition, patients on regular follow-up
579
Clinical Chemistry 45, No. 4, 1999
Table 1. The percentile reference values established using rank and percentile of data for TSH and free and total hormones on the local euthyroid population. TSH fT4 fT3 T4 T3
Units
n
1%
2.5%
5%
mIU/L pmol/L pmol/L nmol/L nmol/L
872 872 146 120 120
0.30 9.6 5.3 59.2 1.49
0.39 10.2 5.6 65.1 1.52
0.54 10.7 5.7 65.5 1.57
with established thyroid disorders were also used in the technical evaluation of the assays. All procedures conformed to the Helsinki Declaration of 1975 and the 1996 revision. Reference intervals were established using rank and percentile confidence limits for euthyroid samples from a multiphasic screening program collected from the local population. The subjects presented with no evidence or clinical suspicion of thyroid abnormalities, including family history of thyroid disorders, pituitary disorders, and psychoses; were not on drugs known to interfere with thyroid gland metabolism or hormone assays; and their results for hepatic and renal function tests and complete blood counts were within the health-related reference intervals (1 ). Table 1 shows the two-tailed 95th interpercentile reference intervals obtained from groups of euthyroid subjects. We measured TSH in 11 pools of measurable hyperthyroid serum (repeated over 5 consecutive days; n 5 9). The TSH value at 20% CV (functional sensitivity) was 0.0032 mIU/L. To provide an efficient first-line test in screening for thyroid disorders in ambulatory patients (2 ), TSH functional sensitivity is important (3 ). As reported previously, a functional sensitivity of 0.0082 mIU/L for TSH was obtained by our laboratory on the DPC Immulite (4 ). At 0.05– 63 mIU/L we found within-day CVs #3.5% (n 5 4) and a total CV #8.4% (n 5 9). The linearity of the TSH assay was determined with nine ratios of two serum samples with concentrations of 129 and 1.08 mIU/L. A linear correlation coefficient of 1.000 (n 5 9) was obtained for TSH. Recoveries were 100 –103%. TSH results obtained from the Vitros ECi were compared with the DPC Immulite and Abbott AxSym assays: ECi 5 0.996 (Immulite) 1 0.161 (n 5 269; r 5 0.993; Syux 5 0.05; P ,0.05); and ECi 5 0.914 (AxSym) 1 0.110 (n 5 147; r 5 0.999; Syux 5 0.02; P ,0.05). Bias plots show that the percent relative difference of TSH between the Immulite or AxSym and the Vitros ECi is spread evenly through the range 0.1–100 mIU/L (Fig. 1). However, at values ,0.1 mIU/L, a positive percent relative difference was observed, indicating that the results obtained on the Vitros ECi were lower than those obtained on either the Immulite or the AxSym. The total imprecision of the free-hormone assays was measured using pooled serum samples of five different concentrations covering the dynamic range for fT4, fT3, T4, and T3. The imprecision analysis was performed on nine occasions over 5 consecutive days. fT4 concentrations
50%
1.35 13.8 6.9 89.3 1.93
95%
97.5%
99%
3.38 17.8 8.3 125.0 2.45
3.89 19.2 8.3 125.0 2.45
4.35 21.7 8.4 126.0 2.58
ranged from 5.1 to 77.4 pmol/L, with a within-day CV #3.2% and a total CV #3.1%. fT3 concentrations ranged from 5.3 to 19.4 pmol/L, with a within-day CV #5.0% and a total CV #4.0%. Total T4 concentrations ranged from 68 to 234 nmol/L, with a within-day CV #1.6% and a total CV #2.8%. Total T3 concentrations ranged from 2.0 to 6.3 nmol/L with a within-day CV #1.3% and a total CV #2.3%. In comparison, our current AxSym analyzer had an interassay CV of 3.5–10% for fT4 at concentrations of 4.0 –77.2 pmol/L (5 ). Browning et al. (6 ) recommend a total CV ,8% for fT4 and ,6% for T4. These targets were achieved in this evaluation. The linearity on dilution was tested for T4 and T3. This was performed by diluting different proportions of a patient sample pool from the upper and lower end of the dynamic range for each analyte to be evaluated. The “recovered” amount of analyte measured for T4 ranged from 88% to 99%, with a correlation coefficient of 0.99. The recovery for T3 ranged from 85% to 95%, with a correlation coefficient of 0.99. We serially diluted patient samples with a protein-free matrix (normal saline 1:2 to 1:16) and assayed for fT4 and fT3. Samples from late pregnancy and critically ill subjects
Fig. 1. Bias plots depicting relative differences of the AxSym and Immulite TSH results and AxSym free and total hormone results with Vitros ECi. % Relative difference (AxSym) 5 [(AxSym 2 ECi)/AxSym] 3 100; % relative difference (Immulite) 5 [(Immulite 2 ECi)/Immulite] 3 100.
580
Technical Briefs
were included. Mean fT4 results were, for samples from euthyroid subjects, 103%, 107%, 109%, and 99% of the undiluted sample; mean fT3 results were, for samples from euthyroid subjects, 100%, 102%, 98%, and 101%; for samples from late-pregnancy subjects, 95%, 95%, 100%, and 109%; for samples from critically ill subjects, 89%, 90%, 92%, and 87%, with three of the groups within 6 10%, as required by Ekins (7 ). On the Vitros ECi analyzer, the fT4 assay uses a patented assay design, including an anti-T4 antibody labeled with horseradish peroxidase, that has been optimized to minimize the effect of binding protein alterations. This fT4 and fT3 assay withstood up to a 1:16 sample dilution (maintaining constancy in the ratio of free hormone to binding proteins) in a protein-free matrix. Albumin is added to the free hormone assay reagent by a number of manufacturers to buffer the effects of increased nonesterified fatty acids that develop in serum in vitro. In our hands, in pregnant subjects, where the equilibrium is altered and the binding capacity is increased, slightly higher fT4 was observed on the Vitros ECi for values .10 pmol/L; values ,10 pmol/L were slightly lower than the AxSym results. Our evaluation data suggested that in critically ill subjects with an altered equilibrium and a low binding capacity, the albumin falsely lower the fT4 concentration reported by the AxSym. In this context, there is no albumin added to the ingredients of the free hormone assays of Vitros ECi. There is also minimal sample:reagent dilution (1:5), a factor that minimizes the underestimation of fT4 when T4 binding inhibitors are present (8 ). These findings in the free hormone assay currently are being investigated further. The fT4, fT3, T4, and T3 results compared with the Abbott AxSym: ECi fT4 5 1.143 (AxSym) 2 3.009 (n 5 338; r 5 0.978; Syux 5 0.12; P ,0.05); ECi fT3 5 1.161 (AxSym) 1 1.723 (n 5 115; r 5 0.932; Syux 5 0.07; P ,0.05); ECi T4 5 0.87 (AxSym) 1 1.327 (n 5 112; r 5 0.983; Syux 5 0.55; P ,0.05); ECi T3 5 1.045 (AxSym) 1 0.174 (n 5 109; r 5 0.961; Syux 5 0.02; P ,0.05).
The bias plots of the free and total hormones are shown in Fig. 1. T3 and fT3 results on the Vitros ECi were significantly higher (fT3 ranging from 20% to 150% higher; T3 up to 50% higher at lower concentrations) than the AxSym. T4 results reported on the Vitros ECi are 10 –20% lower than AxSym results. fT4 values ,10 pmol/L on the AxSym appear to be higher. The lower values on the Vitros ECi may be considered more valid because there is no albumin in the assay reagent to buffer the effects of nonesterified fatty acids (7 ). In conclusion, the technical performance, ease of operation, and rapid turnaround time makes the Vitros ECi a consideration for routine use in thyroid evaluation.
This work was supported by Ortho Clinical Diagnostics, Singapore, who kindly supplied all the test kits used in the analyses. We thank the staff of the Department of Laboratory Medicine, National University Hospital, Singapore, for technical assistance. References 1. International Federation of Clinical Chemistry. Approved recommendation (1987) on the theory of reference values. Part 2. Selection of individuals for the production of reference values. J Clin Chem Clin Biochem 1987;25: 639 – 44. 2. Wilkinson E, Rae PWH, Thomson KJT, Toft AD, Spencer CA, Beckett GJ. Chemiluminescent third-generation assay (Amerlite TSH-30) of thyroidstimulating hormone in serum or plasma assessed. Clin Chem 1993;39: 2166 –73. 3. Nelson JC, Wilcox RB. Analytical performance of free and total thyroxine assays. Clin Chem 1996;42:146 –54. 4. Aw TC, Ling E. Performance of a third generation TSH assay on the random-access analyzer Immulite [Abstract]. Clin Chem 1994;40:1029 –30. 5. Aw TC, Ling E. Determinations of free thyroxine (fT4) on automated analysers [Abstract]. Proc 10th Asia-Ocean Congr Endocrinol, 1994:380. 6. Browning MCK, Ford RP, Callaghan SJ, Fraser CG. Intra- and interindividual biological variation of five analytes used in assessing thyroid function: implications for necessary standards of performance and the interpretation of results. Clin Chem 1986;32:962– 6. 7. Ekins R. Measurement of free hormones in blood. Endocr Rev 1990;11:5– 46. 8. Nelson JC, Weiss RM. The effect of serum dilution on free thyroxine (T4) concentration in the low T4 syndrome of nonthyroidal illness. J Clin Endocrinol Metab 1985;61:239 – 46.
