Context.-It is important that the total long-term precision of laboratory methods meet the medical needs of the patients being served.

Objectives.-To determine the long-term within- and between-laboratory variation of cortisol, ferritin, thyroxine, free thyroxine, and thyroid-stimulating hormone measurements using commonly available methods and to determine if these variations are within accepted medical needs.

Design.-Two vials of pooled frozen serum were mailed 6 months apart to laboratories participating in 2 separate College of American Pathologists surveys. The data from those laboratories that analyzed an analyte in both surveys were used to determine for each method the total variance and the within- and between-laboratory components.

Setting.-The study included the A mailing of the 2003 College of American Pathologists Ligand Survey and the C mailing of the Chemistry Survey.

Main Outcome Measures.-For each analyte, total variance was partitioned into within- and between-laboratory components for each analytic method. The within-laboratory variations were then compared with imprecision criteria based on biological variation.

Participants.-The laboratories that reported results on the same analyte using the same method in both surveys.

Results.-For each analyte, the median of the long-term within-laboratory variances of each peer group was 78% to 95% of its total-survey variance, and the median long-term within-laboratory coefficients of variation varied from 5.1 % to 7.6%. The number of methods that met within-laboratory imprecision goals based on biological criteria were 5 of 5 for cortisol; 5 of 7 for ferritin; 0 of 7 for thyroxine and free thyroxine; and 8 of 8 for thyroid-stimulating hormone.

Conclusions.-For all analytes tested, the total within-laboratory component of variance was the major source of variability in this study. In addition, there are several methods, especially for thyroxine and free thyroxine, that may not meet analytic goals in terms of their imprecision.

(Arch Pathol Lab Med. 2005;129:318-322)

Since 1974, College of American Pathologists (CAP) surveys have been utilized to study the sources of variation of clinical laboratory tests. Several of these studies1-4 have reported long-term variability using different types of replicate proficiency testing material sent out several months apart. Unfortunately, it was not known if proficiency testing materials accurately reflected clinical material with respect to long-term imprecision.

In 1994, CAP implemented a fresh frozen sera study hoping to better understand methodologic biases and matrix effects.5 In that study, a pool of sera was frozen in aliquots, which were mailed as part of the Chemistry Survey. In 2003, a similar study was done to examine both common chemistry and endocrine analytes. Because 1 aliquot of the same frozen serum pool was mailed in both the Chemistry and Ligand surveys, there were replicate samples sent during a period of 6 months to many laboratories. This provided an opportunity to examine the total (long-term) within-laboratory precision of the common methods for several tests using material that represents real patient samples. Although several CAP studies1-4 have examined the within-laboratory and between-laboratory variability components of the total variance, none had used the fresh serum specimen. The purpose of this study was to determine the long-term within-laboratory variation of cortisol, ferritin, thyroxine, free thyroxine, and thyroid-stimulating hormone measurements using fresh frozen serum replicate samples for commonly available methods and to determine if these variations are within accepted medical needs.

MATERIALS AND METHODS

Samples and Study Design

The preparation and the chemical characteristics of the fresh frozen serum samples have been described previously.6 In the A mailing of the 2003 CAP Ligand Survey, an aliquot of the fresh frozen serum pool was sent as 1 of the 5 challenges. Approximately 6 months later, an aliquot of the same fresh frozen serum pool was included in the C mailing of the 2003 Chemistry Survey. Both samples were analyzed by the survey participants in the manner prescribed by the Clinical Laboratory Improvement Amendments of 1988.7

Data

The results obtained for the fresh frozen serum samples were analyzed and reported back to the participants in the manner typical of the respective surveys. The data were then re-analyzed in a manner similar to a previous study involving therapeutic drugs.1 That is, the data were run through a 2-pass, 3-SD screening program to remove outliers. Only those laboratories participating in both surveys using the same analytic method were included. In addition, those data from methods not specified or from methods with fewer than 20 participants were not included, because these data tended to produce unreliable variance distributions. The analytic methods studied are listed in Table 1.

RESULTS

Tables 2 and 3 show the summary of the study results. For each analyte, the median total within-laboratory variance ranged from 78.4% to 95.2% of the total-survey variance. The lowest percentage of within-laboratory variance for any individual method was 64.1% (ferritin). In addition, the least precise method for an analyte typically had within-laboratory and total-survey coefficients of variation that were approximately 2 to 3.5 times that of the most precise method.