Steele, Bernard W.

Abstract Context. —In proficiency testing surveys, there are differences in the values reported by users of various analytic methods. Two contributors to this variation are calibrator bias and matrix effects of proficiency testing materials. Objectives. —(1) To quantify the biases of the analytic methods used to measure thyroid-stimulating hormone, thyroxine, triiodothyronine, free thyroxine, and free triiodothyronine levels; (2) to determine if these biases are within allowable limits; and (3) to ascertain if proficiency testing materials correctly identify these biases. Design. —A fresh frozen serum specimen was mailed as part of the 2003 College of American Pathologists Ligand and Chemistry surveys. The means and SDs for each analytic method were determined for this sample as well as for a proficiency testing sample from both surveys. In the fresh frozen serum sample, target values for thyroxine and triiodothyronine were determined by isotope dilution/liquid chromatography/tandem mass spectrometry. All other target values in the study were the median of the means obtained for the various analytic methods. Main Outcome Measures. —Calibration biases were calculated by comparing the mean of each analytic method with the appropriate target values. These biases were evaluated against limits based on intra- and interindividual biological variation. Matrix effects of proficiency testing materials were assessed by comparing the rank of highest to lowest analytic method means (Spearman rank test) for each analyte. Participants. —Approximately 3900 clinical laboratories were enrolled in the College of American Pathologists Chemistry and Ligand surveys. Results. —The number of methods in the Ligand Survey that failed to meet the goals for bias was 7 of 17 for thyroid-stimulating hormone and 11 of 13 for free thyroxine. The failure rates were 12 of 16 methods for thyroxine, 8 of 11 for triiodothyronine, and 9 of 11 for free triiodothyronine. The means of the analytic method for the proficiency testing material correlated significantly ( P < .05) only with the fresh frozen serum means for thyroxine and thyroid-stimulating hormone in the Chemistry Survey and free triiodothyronine in the Ligand Survey. Conclusions. —A majority of the methods used in thyroid function testing have biases that limit their clinical utility. Traditional proficiency testing materials do not adequately reflect these biases.

Abstract Objective.—To identify the sources of analytical variation for cyclosporine and tacrolimus in a 3-year longitudinal study. Design.—Two pools of whole blood were spiked with cyclosporine and tacrolimus, respectively. One aliquot of cyclosporine and 2 of the tacrolimus pool were distributed in the first and last mailing for years 1999 to 2001. For both drugs, the total variance for each method was partitioned into within- and between-laboratory components. Setting.—The A and C mailings of the 1999, 2000, and 2001 AACC/CAP [American Association for Clinical Chemistry/College of American Pathologists] Immunosuppressive Drugs (CS) Monitoring Survey. Main Outcome Measures.—For each drug, total variance was partitioned into specimen, mailing, year, and interlaboratory effects for each analytical method. Participants.—The 292 laboratories for cyclosporine and 177 laboratories for tacrolimus enrolled in the survey from 1999 to 2001. Results.—For both cyclosporine and tacrolimus, the major source of imprecision came from within-laboratory factors, which accounted for nearly 85% (range, 77% to 90%) of the total variance. For cyclosporine, the major component of within-laboratory variance was between-mailing, within-year effect, whereas for tacrolimus it was the between-year, within-laboratory variation. Conclusion.—The major source of long-term survey imprecision for cyclosporine and tacrolimus is within-laboratory factors. The finding that 85% of the total variance was due to within-laboratory variation is similar to other therapeutic drugs.

Abstract Objective.—To determine if the levels of imprecision of the commonly used analytic methods for drug measurements are suitable for long-term therapeutic drug monitoring. Design.—In 1996, 4 identical lyophilized samples (2 in the first mailing and 2 in the second mailing 4 months later) were sent to laboratories participating in a nationwide proficiency testing program. Similarly, in 1999, replicates from a liquid pool of spiked sera were mailed 3 times, 4 months apart, to participating laboratories. For each of 11 drugs regulated under the Clinical Laboratory Improvement Amendments of 1988 and 1 metabolite, the total variance for each method was partitioned into within- and between-laboratory components. The total within-laboratory and the total survey coefficients of variation (CVs) for each method were then compared with the “acceptable” precision criteria of Glick, Burnett, and Fraser for each drug. Setting.—The first 2 mailings of the College of American Pathologists Therapeutic Drug Monitoring surveys for 1996, sets Z and ZM, and the 3 mailings of 1999, sets ZM, Z, and Z2. Main Outcome Measures.—For each drug studied, the CV of each method was compared with the various imprecision criteria, and if greater than any of the criteria, the method was then tabulated as not meeting that specific criterion. Participants.—The approximately 5000 participants of the survey. Results.—The number of methods deemed as not having acceptable total long-term within-laboratory precision by the various criteria ranged from 35% to 88% in 1996 and from 22% to 77% in 1999. Conclusion.—The number of failures possibly indicates that many of the reagent assays being utilized are not precise enough for long-term therapeutic drug monitoring of chronically administered drugs or that the published criteria used to evaluate the data in this study are too stringent.

Abstract Objective. —To determine the magnitudes and sources of analytic variation in testing for therapeutic drugs. Specifically, among laboratories using the same analytic method, to compare the within-laboratory variation (including both short- and long-term variation) with the between-laboratory variation. Design. —Four identical challenges were prepared from a lyophilized pool of spiked sera and were sent in pairs 4 months apart to laboratories participating in a nationwide proficiency-testing program. For each of 25 drugs, the variability in reported results from laboratories using the same method was investigated using nested analysis of variance. Setting. —The first 2 mailings of the College of American Pathologists Therapeutic Drug Monitoring Survey, 1996, sets Z and ZM. Main Outcome Measures. —For each drug, total variance was partitioned into within- and between-laboratory components for common methods. The within-laboratory component was further partitioned into short- and long-term components. Participants. —The approximately 5000 laboratories enrolled in the survey. Results. —For the 25 drugs, the average percentages of the total variance due to short-term, within-laboratory variance; long-term, within-laboratory variance; between-laboratory variance; and total laboratory variance were 25.0% (range, 8.8–50.6%), 57.8% (35.3–73.7%), 17.3% (5.0–35.4%), and 82.7% (64.6–95.0%), respectively. Conclusion. —For all drugs tested, the within-laboratory component of variance was greater than the between-laboratory component of variance. Within laboratories, the magnitude of the long-term component was generally greater than the magnitude of the short-term component. This information will be helpful in determining the clinical utility of various drug assays and in evaluating the appropriateness of regulations involving therapeutic drug testing.