Printer Friendly

Laboratory performance in albumin and total protein measurement using a commutable specimen: results of a college of American pathologists study.

Discrepant results for serum constituents observed among peer groups in the College of American Pathologists (CAP) Comprehensive Chemistry Survey may be caused by differences in calibration and method specificity for the analyte or by noncommutability (matrix-related bias) of the survey specimens with clinical patient samples. Noncommutability is caused by changes in the physical and chemical characteristics introduced by the manufacturing process used to prepare survey specimens. (1,2)

To examine biases among peer groups, the CAP Chemistry Resource Committee included a one-time, off-the-clot, human, fresh frozen serum (FFS) specimen in the 2003 Fall Comprehensive Chemistry Survey. By virtue of the process used for its preparation, this specimen was assumed to be commutable or to exhibit no matrix-related bias among the various analytic methods used by the survey participants. (3) Consequently, methods having good chemical specificity and proper calibration would be expected to produce the same result for a given analyte for the FFS specimen. In this communication, we present and discuss the results reported by participants for albumin and total protein.

MATERIALS AND METHODS

Survey specimen C-02 was frozen aliquots from a freshly collected, off-the-clot, human serum pool. A detailed description of the preparation of the C-02 specimen and the 4 conventional survey specimens, along with information on survey logistics, has been previously reported. (3) The protein base of the conventional specimens was defibrinated human plasma to which various analyte concentrates (but no bovine proteins) were added. Specimen C-02 was assumed to be commutable with native patient samples based on the stringency of its preparation protocol. This protocol has been validated to produce commutable specimens for total cholestero (l,4,5) high-density lipoprotein-cholesterol, triglycerides and apolipoproteins, (5) and creatinine. (6)

We assessed the commutability of the 4 conventional proficiency testing (PT) specimens with specimen C-02 from the differences between the instrument groups mean values for the conventional specimens and the differences for the C-02 specimen from the reference or the designated comparison measurement procedure values.

Human serum albumin, Cohn fraction V, solution, 10 g/dL, was purchased from MP Biomedicals (Irvine, California); bromocresol green (BCG) sodium salt and bromocresol purple (BCP) free acid were obtained from JT Baker Company (Phillipsburg, New Jersey) and Fluka Company (St. Louis, Missouri), respectively. Bovine serum albumin (7% Solution), standard reference material 927c, was obtained from the National Institute of Standards and Technology (Gaithersburg, Maryland).

Albumin was measured by 2 dye-binding methods, designated as comparative methods, in 2 reference laboratories. Analyses by the BCG method (7) were performed at the reference standards laboratories of Children's Hospital of Wisconsin (CHW, Milwaukee, Wisconsin) and at Ortho Clinical Diagnostics (Rochester, New York). To improve the specificity of the assay for albumin, absorbance measurements were made 30 seconds after adding the sample to the BCG reagent (at CHW) and after 6 seconds at Ortho Clinical Diagnostics. (8) Specimens C-01, C-02, and C-05 were analyzed in triplicate at CHW and Ortho Clinical Diagnostics in April 2003. All 5 specimens were analyzed in duplicate at CHW in May 2004. For calibration, both laboratories used a 10 g/dL human serum albumin solution, the albumin purity of which, according to the manufacturer's certificate, was 96%. The albumin concentration assigned to this solution was determined as the total protein measured by the biuret reference method. (9) Working calibrators, containing from 1 to 5 g/dL albumin (nominal values), were prepared by diluting the 10 g/dL human serum albumin solution with 0.9 g/dL NaCl and 0.05 g/dL sodium azide in water. Analyses by the BCP method were performed at CHW by the Pinnell and Northam (10) procedure. To improve the linearity of the assay, the concentration of BCP in the working reagent was increased from 40 to 60 umol/L. The procedure was calibrated with the human serum albumin calibrators used with the BCG method.

Total protein was measured by the biuret reference measurement procedure. (9) Specimens C-01, C-02, and C-05 were analyzed in triplicate at CHW, Ortho Clinical Diagnostics, Canadian External Quality Assessment Laboratory (Vancouver, British Columbia, Canada), and Wisconsin State Laboratory of Hygiene (Madison, Wisconsin) in April 2003. Specimens C-03 and C-04 were analyzed in triplicate at Wisconsin State Laboratory of Hygiene (in April 2003) and at CHW (June 2004). All total protein assays were calibrated with standard reference material 927c bovine serum albumin (7% solution) from the National Institute of Standards and Technology.

Data were collected via the CAP survey result reporting system. Outliers were removed based on the usual CAP statistical analysis with a 2-pass, 3-SD criteria. (3) After outlier removal, results from peer groups with fewer than 10 participants were removed. Peer-group mean values were compared with the mean values determined by the appropriate designated comparison or reference methods with a 2-sample, 2-sided t test. All statistical analysis was performed with SAS for Windows software (version 8.2, SAS Inc, Cary, North Carolina) or Excel 2000 (Microsoft Corporation, Redmond, Washington).

RESULTS

Albumin

In the absence of a reference measurement procedure and a certified reference material for albumin, well-characterized BCG and BCP procedures were used as designated comparison methods. Table 1 shows peer-group mean values for albumin measured by BCG or BCP methods for the FFS C-02 reported by all participants in the survey. In addition, the standard deviations, standard errors of the mean, and differences between the peer group means and the mean values obtained by the designated comparison methods are presented. Mean values were in apparent agreement for all methods between the BCG and BCP procedures (4.07 g/dL versus 4.06 g/dL). The apparent agreement was due to the much lower mean produced by the VITROS instruments (3.82 g/dL; Ortho Clinical Diagnostics) than that produced by all other peer groups (4.27 g/ dL) that used the BCG method and because the VITROS instruments accounted for 45% of all data reported using the BCG method. It would be statistically inappropriate to include the VITROS results in calculating the mean from all BCG methods when the difference between the VITROS values and the means from all other methods was statistically significant (P < .001).

For the BCG methods, 15 of 21 peer-group mean values (71%) were significantly different (P < .001), and for the BCP methods, all (100%) of peer group means were significantly different (P < .001) from the mean values of the respective, designated comparison method values. For all BCG results, except VITROS, the average difference was 0.15 g/dL. The Beckman-Coulter Synchron systems (Brea, California) had mean values close to those of the comparative method. The other peer groups had differences that ranged from 0.10 g/dL to 0.32 g/dL. All BCP methods (100%) had 0.25 to 0.47 g/dL positive differences from the comparative method.

For the subsequent presentation of results, similar instruments and methods from the same manufacturer that had similar means and standard deviations are grouped into a single category. For example, Hitachi" results for BCG methods comprise the combination of 6 different Hitachi (Indianapolis, Indiana) instruments, including the Roche (Indianapolis, Indiana) Modular instrument. The Toshiba BCG (Tokyo, Japan) method was excluded from further assessment because of the few participants.

Table 2 shows summary statistics for aggregated instrument groups. For the BCG methods, standard deviations for the 5 survey specimens and the 6 instrument groups ranged from 0.05 g/dL (specimen C-03, ADVIA, Tarrytown, New York) to 0.15 g/dL (specimen C-02, VITROS). Standard deviations for the BCP methods for the same survey specimens and the 5 instrument groups were slightly smaller (0.04-0.11 g/dL) than were those for the BCG methods.

Total Protein

Table 3 shows peer-group mean values, standard deviations, standard errors of the mean, and biases for total protein for specimen C-02 reported by all participants in the survey. Mean values for 12 of 24 peer groups (50%) are statistically different (P <.001) from the values of the reference-measurement procedure. However, the biases of the mean values for all but one peer group were small, from--0.07 to 0.15 g/dL. The Beckman Synchron LX20 had a bias--0.30 g/dL from the value of the reference measurement procedure.

For the subsequent presentation of results, similar instruments from the same manufacturer were grouped into a single category, with the Synchron LX20 kept as a separate group because of its larger bias compared with other Beckman instruments. Summary statistics for total protein are shown in Table 4; these instrument groups comprise 97% of the data submitted by participants. Mean values for "All Methods" results were within 60.08 g/dL of those obtained with the reference measurement procedure for all 5 survey specimens. Mean values of the various instrument groups were, with the exception of Synchron LX20, within 60.15 g/ dL of the reference measurement procedure. The mean values for the Synchron LX20 results were 0.15 to 0.32 g/dL lower than they were for the other Synchron versions and 0.15 to 0.34 g/dL lower than they were for the reference measurement procedure. It appears that the application of the biuret method to the Synchron LX20 was different from the other versions of this clinical analyzer.

Method Performance

Differences between albumin mean values and the comparative methods for BCG albumin methods are shown in Figure 1. Four of the instrument groups had positive biases, the values from Synchron were accurate, and those from VITROS were negatively biased. The biases for BCP albumin methods were all positive, were about twice as large as those of the BCG method, and were clustered less tightly (Figure 2). Because many reference intervals for serum albumin provided by instrument manufacturers are almost the same for both BCG and BCP methods, the generally larger positive biases for BCP methods may lead to errors in diagnosis and treatment.

Figure 3 shows the biases for total protein between mean values for instrument groups and those for the reference measurement procedure. Biases for the Synchron LX20 were proportional to the protein concentrations of the specimens. Biases for Synchron LX20 and VITROS varied more among specimens than did the other instrument groups.

COMMENT

Albumin

The Clinical Laboratory Improvement Amendments of 1988 (CLIA) specifies that albumin values within [+ or -] 10% of the peer-group mean value in a PT program represent acceptable performance; thus, the total allowable error corresponds to a coefficient of variation of approximately 5%, assuming no bias because the peer group mean is used as the target value. For specimen C-02 (Table 1), the coefficient of variation for all results (except the VITROS) for the BCG albumin methods was (mean [SD]) 3.3% (0.14) and for the BCP methods was 2.2% (0.09). For the conventional specimens, coefficients of variation for the BCG and BCP methods varied from 3.0% to 4.1% and from 2.3% to 2.8%, respectively. Based on these coefficient of variation values, it would be expected that 95% of the reported results would satisfy CLIA requirements for both method types. However, peer-group grading does not address the trueness or accuracy of the results. Specifications for desirable imprecision rates and biases based on biological variability are 1.6% (one-half of the intra-individual biologic variation) for imprecision, 1.3% for bias, and 4.7% for total error with a coverage factor of 2 (which gives a value similar to a 95% or a [+ or -] 2-SD interval in the absence of bias). (11) Based on biologic-variability criteria, neither BCG nor BCP albumin methods had satisfactory performance based only on coefficient of variation values. The CLIA criteria consider only imprecision and appear to be more reasonable at the present state of the art, which lacks a consensus reference method and a reference material for serum albumin.

Because the traceability of BCG and BCP methods is not known, the term bias is used here, as in other reports, to denote the difference in albumin values between the routine methods and the designated comparison methods. Within these limitations, the results showed that BCG methods produced results that were generally closer to those of the comparative method (Figure 1) than did instruments using BCP methods (Figure 2). However, when considering the bias versus the respective designated comparison methods, neither method type had satisfactory total error performance.

The lack of a reference material and a reference measurement procedure for serum albumin does not allow a definitive statement about accuracy of the various methods. Although we do not know what kinds of calibrators were used by instrument manufacturers, it appeared that calibration had greater uniformity among the BCP methods than was seen among the BCG methods. For specimen C-02, the average difference among all BCP methods was 0.22 g/dL versus 0.39 g/dL for the BCG methods (with VITROS results excluded; Table 1). The calibration of VITROS systems was clearly different from that of the other manufacturers. When comparing the mean values of all methods for BCG (excluding VITROS) to BCP, we observed that the BCP methods were 0.21 g/dL lower than were the BCG methods (Table 1). Other reports (12-15) have documented that albumin values for BCP methods are 0.3 to 0.5 g/dL lower than are those obtained by BCG methods. When comparing our results to the respective designated comparison methods, we observed positive biases of 0 to 0.3 g/dL for BCG methods (excluding VITROS) and larger positive biases of 0.3 to 0.5 for BCP methods. These observations suggest that manufacturers may have calibrated BCP methods with the intention of compensating for the differences in specificity between the 2 dyes for human serum albumin. In our opinion, that kind of harmonization is undesirable because the reference intervals for serum albumin by the BCP and BCG methods are different.

Blagg et al (16) measured serum albumin by BCG and BCP methods in 121 healthy subjects and in 235 unselected patients with hemodialysis. In the healthy subjects, albumin mean values were 4.4 g/dL by BCG methods and were 3.9 g/ dL by BCP methods. In patients with hemodialysis, albumin mean values were 3.8 g/dL by BCG methods and 3.3 g/dL by BCP methods. Because of the difference in albumin values between BCG and BCP methods, the authors recommended the following (16):

Dialysis facilities must be aware of the albumin method used by its clinical laboratory and, based on the particular method, determine appropriate assurance limits. If serum albumin concentration is to be used as a guide to the probability of mortality or of the occurrence of various complications, whether in dialysis patients or other patients, the specific albumin method used must be taken into consideration.

Similar recommendations were issued by the Medical Review Committee of the Renal Network of the Upper Midwest, Inc (Saint Paul, Minnesota) in their report on serum albumin to Network 11 Dialysis Facility Medical Directors on December 6, 1993, to wit (17):

(1) Dialysis facility personnel should become informed about their lab's albumin assay, calibration standards, and reference range. (2) Reference ranges for the serum albumin assay should be based on a normal" (eg, not ESRD--end stage renal disease--or hospitalized patients) population. (3) The goal for adult patients with hemodialysis should be to maintain the serum albumin level within the lab's reference range.

Clinical Significance and Utility of Serum Albumin Measurements

The clinical utility of albumin measurements is well established. Lowrie et al (18) were the first to report that in 12 000 patients with hemodialysis low serum albumin was strongly associated with a risk for death. The risk of death in patients with albumin values (measured by BCG) between 4.01 to 4.50 g/dL was 9.7%; it increased to 21.5% with albumin values of 3.51 to 4.00 g/dL, to 47.2% with albumin values of 3.01 to 3.50 g/dL, and to 68% in patients with albumin values of 2.51 to 3.00 g/dL. Owen et al (19) confirmed the Lowrie et al (18) observations and found that low serum albumin concentration (<4 g/dL by BCG) was a more powerful (21 times stronger) predictor of risk of death than the urea reduction ratio. Hermann et al (20) found that patients hospitalized for acute illness with albumin concentrations (by BCG) of less than 3.4 g/dL, who made up 21% of the population of 15 500 studied, were more likely to die, had longer hospital stays, and were readmitted sooner and more frequently than patients with albumin levels within reference range. The in-hospital mortality was 4% with albumin levels greater than 3.4 g/dL and 14% with albumin levels less than 3.4 g/dL. Similar results were reported by other investigators. (21,22) In a recent publication, Friedman and Fadem (23) stressed that low serum albumin be recognized as a clinical index of illness rather than of malnutrition. Because there is a strong inverse relationship between hypoalbuminemia and mortality, low albumin in patients with chronic kidney disease should prompt physicians to identify and treat underlying disorders.

Our data suggest that guidelines for interpreting serum albumin are applicable only for the methods used to develop the guidelines and should not be generalized to interpreting results from other serum albumin methods.

In view of the importance of serum albumin as an independent predictor of risk of morbidity and mortality, we add our support to previous recommendations urging clinical laboratories and instrument manufacturers to establish reference intervals for serum albumin by the methods used with their instruments. Guidelines for interpretation of albumin results should be based on method-specific reference intervals until albumin methods are standardized to produce comparable results from different methods.

The development of a certified reference material or, at minimum a "consensus reference material," for human albumin is important to enabling the establishment of traceability to a higher-order reference system to improve accuracy in the measurement of albumin.

Total Protein

An acceptable performance result in CLIA for total protein is 610% of the peer-group mean value. Specifications based on biologic variability for desirable imprecision rates, biases, and total errors are 1.4%, 1.2%, and 4.1% (with coverage factor 2), respectively. (11) Based primarily on the C-02 results, both of these requirements were satisfied by the instrument systems represented in this survey, except by the Beckman Synchron LX20, which appeared to have a relatively large calibration bias. Although all but 11 of the 45 peer-group mean values (76%) were significantly different (P < .001) from those of the reference method, the differences were generally too small to be of any clinical significance. The relatively larger biases of the protein values reported by the Synchron LX20 users appeared to be proportionally related to the protein concentration of the specimens (Figure 3).

Commutability of PT Materials

A survey PT specimen is considered commutable when the mathematic relationships between the results from different measurement procedures are equivalent for the survey specimens and for the clinical patient samples. Table 5 shows the differences between the mean values for the instrument groups with conventional PT specimens and the differences in values for the C-02 (FFS) specimen from the reference or designated comparison measurement procedure. A conventional PT specimen was considered commutable with the C-02 FFS specimen if its difference was within [+ or -] 2 SD of the difference for the C-02 specimen for the instrument group. The commutable specimen C-02 included in this survey allowed assessment of the state of the art for accuracy of individual laboratories, the standardization and trueness versus a reference measurement procedure for routine total protein methods, and the state of harmonization or agreement among routine albumin test methods as well as their bias versus robust, designated comparison methods. The design also allowed determination of the commutability of the conventional PT specimens compared with the C-02 specimen.

For routine albumin methods, 23 of 24 BCG method-materials (96%) and 13 of 16 BCP method-material combinations (81%) met the commutability criterion (Table 5). For the total protein methods, all 36 method-material combinations (100%) for the conventional survey specimens met the commutability criterion. Because each new lot of PT specimens may have had different characteristics and because different reagent lots can affect commutability characteristics, our commutability observations cannot be generalized to other preparations of conventional PT specimens. (24) Some of the results for the noncommutable samples could have suggested incorrect conclusions about method performance. Because the commutability characteristic of a given PT specimen is generally unknown, its results must be interpreted with caution to avoid incorrect conclusions regarding a method's performance.

Caption: Figure 1. Difference between instrument-group mean and that of the designated comparison method for albumin measured by bromocresol green methods for College of American Pathologists survey specimens. The point for specimen C-05 is missing for the VITROS instrument because that specimen was noncommutable. Key: ^, C-01; u, C-02 (fresh frozen serum); n, C-03; X, C-04; *, C-05. Instrument companies: ADVIA (Siemens), Tarrytown, New York; Cobas Integra (Roche), Indianapolis, Indiana; Hitachi (Roche), Indianapolis, Indiana; Olympus (Beckman), Brea, California; Synchron (Beckman), Brea, California; VITROS (Ortho Clinical Diagnostics), Rochester, New York.

Caption: Figure 2. Difference between instrument-group mean and the designated comparison method for albumin measured by bromocresol purple methods for College of American Pathologists survey specimens. The points for specimens C-03 and C-04 are missing for the Aeroset and the points for specimen C-05 are missing for the Synchron because those specimens were noncommutable. Key: ^, C-01; u, C-02 (fresh frozen serum); n, C- 03; X, C-04; *, C-05. Instrument companies: Aeroset (Abbott), Chicago, Illinois; Dimension (Siemens), Tarrytown, New York; Hitachi (Roche), Indianapolis, Indiana; Synchron (Beckman), Brea, California.

Caption: Figure 3. Difference between instrument-group mean and the reference method procedure for total protein in College of American Pathologists survey specimens. Key: ^, C-01; u, C-02 (fresh frozen serum); n, C-03; X, C-04; *, C-05. Instrument companies: ADVIA (Siemens), Tarrytown, New York; Aeroset (Abbott), Chicago, Illinois; Cobas Integra (Roche), Indianapolis, Indiana; Dimension (Siemens), Tarrytown, New York; Hitachi (Roche), Indianapolis, Indiana; Olympus (Beckman), Brea, California; Synchron (Beckman), Brea, California; VITROS (Ortho Clinical Diagnostics), Rochester, New York.

References

(1.) Miller WG. Specimen materials, target values and commutability for external quality assessment (proficiency testing) schemes. Clin Chim Acta. 2003; 327(1-2):25-37.

(2.) Miller WG, Jones GRD, Horowitz GL, Weykamp C. Proficiency testing/ external quality assessment: current challenges and future directions. Clin Chem. 2011;57(12):1670-1680.

(3.) Miller WG, Myers GL, Ashwood ER, et al. Creatinine measurement: state of the art in accuracy and interlaboratory harmonization. Arch Pathol Lab Med. 2005;129(3):297-304.

(4.) Clinical and Laboratory Standards Institute. Preparation and Validation of Commutable Frozen Human Serum Pools as Secondary Reference Materials for Cholesterol Measurement Procedures; Approved Guideline. Wayne, PA: CLSI; 1999. CLSI publication C37-A; vol 19, no. 25.

(5.) Cobbaert C, Weykamp C, Baadenhuijsen H, Kuypers A, Lindemans J, Jansen R. Selection, preparation, and characterization of commutable frozen human serum pools as potential secondary reference materials for lipid and apolipoprotein measurements: study within the framework of the Dutch project "Calibration 2000." Clin Chem. 2002;48(9):1526-1538.

(6.) National Kidney Disease Education Program, US Department of Health and Human Services. Commutability study of creatinine reference materials. http://nkdep.nih.gov/lab-evaluation/gfr/creatinine-standardization/ commutability-study.shtml. Updated March 1, 2012. Accessed July 11, 2012.

(7.) Doumas BT, Watson WA, Biggs HG. Albumin standards and the measurement of serum albumin by bromcresol green. Clin Chim Acta. 1971; 31(1):87-96.

(8.) Gustafsson JEC. Automated serum albumin determination by use of the immediate reaction with bromcresol reagent. Clin Chem. 1978;24(2):369-373.

(9.) Doumas BT, Bayse DD, Carter RJ, Peters T Jr, Schaffer R. A candidate reference method for determination of total protein in serum, I: development and validation. Clin Chem. 1981;27(10):1642-1650.

(10.) Pinnell AE, Northam BE. New automated dye-binding method for serum albumin determination with bromcresol purple. Clin Chem. 1978;24(1):80-86.

(11.) Ricos C, Alvarez V, Cava F, et al. Current databases on biologic variation: pros, cons, and progress. Scand J Clin Lab Invest. 1999;59(7):491-500.

(12.) Webster D, Bignell AHC, Atwood EC. An assessment of suitability of bromocresol green for the determination of serum albumin. Clin Chim Acta. 1974;53(1):101-108.

(13.) Webster D. The immediate reaction between bromocresol green and serum as a measure of albumin content. Clin Chem. 1977;23(4):663-664.

(14.) Speicher CE, Widish JR, Gaudot FJ, Hepler BR. An evaluation of the overestimation of "acute phase reactants" by use of the bromcresol green reaction. Am J Clin Pathol. 1978;69(3):347-350.

(15.) Gustafsson JEC. Improved specificity of serum albumin determination and estimation of "acute phase reactants" by use of the bromcresol green reaction. Clin Chem. 1976;22(5):616-622.

(16.) Blagg CR, Liedtke RJR, Batjer JD, et al. Serum albumin concentration-related Health Care Financing Administration quality assurance criteria is method dependent: revision is necessary. Am J Kidney Dis. 1993;21(2):138-144.

(17.) Medical Review Committee. Notice to Network 11 Facilities Regarding Albumin Results in the Quality Improvement Demonstrations. St. Paul, Minnesota: Renal Network of the Upper Midwest, Inc; 1993.

(18.) Lowrie EG, Lew NL. Death risk in hemodialysis patients: the predictive value of commonly measured variables and an evaluation of death rate differences between facilities. Am J Kidney Dis. 1990;15(5):458-482.

(19.) Owen WF Jr, Lew NL, Liu Y, Lowrie EG, Lazarus JM. The urea reduction ratio and serum albumin concentration as predictors of mortality in patients undergoing hemodialysis. NEngl J Med. 1993;329(14):1001-1006.

(20.) Hermann FR, Safran C, Levkoff SE, Minaker KL. Serum Albumin level on admission as a predictor of death, length of stay, and readmission. Arch Intern Med. 1992;152(1):125-130.

(21.) Corti M-C, Guralnik JM, Salive ME, Sorkin JD. Serum albumin level and physical disability as predictors of mortality in older persons. JAMA. 1994; 272(13):1036-1042.

(22.) Guijarro C, Massey ZA, Wiederkehr MR, Ma JZ, Kasiske BL. Serum albumin and mortality after renal transplantation. Am J Kidney Dis. 1996;27(1): 117-123.

(23.) Friedman AN, Fadem SZ. Reassessment of albumin as a nutritional marker in kidney disease. J Am Soc Nephrol. 2010;21(2):223-230.

(24.) Miller WG, Erek A, Cunningham TD, Oledipo O, Scott MG, Johnson RE. Commutability limitations influence quality control results with different reagent lots. Clin Chem. 2011;57(1):76-83.

Stanley F. Lo, PhD; W. Greg Miller, PhD; Basil T. Doumas, PhD

Accepted for publication August 23, 2012.

From the Reference Standards Laboratory, Department of Pathology, Children's Hospital of Wisconsin, Milwaukee (Drs Lo and Doumas);and the Department of Pathology, Virginia Commonwealth University, Richmond (Dr Miller).

The authors have no relevant financial interest in the products or companies described in this article.

Reprints: Stanley F. Lo, PhD, Reference Standards Laboratory, Department of Pathology and Laboratory Medicine, Children's Hospital of Wisconsin, 9000 W Wisconsin Ave, Milwaukee, WI 53226 (e-mail: slo@mcw.edu).

Table 1. Statistical Data for Peer Groups for Albumin Specimen C-02
(Fresh Frozen Serum) by the Bromocresol Green (BCG) and Bromocresol
Purple (BCP) Methods

                                                   Albumin Results
                                                   by BCG, (a) g/Dl

                                   Laboratories,       C-02
         Instrument (b)               No. (%)      Mean     SD

Designated comparison                              4.12
  method assigned value
Bayer ADVIA 1650                     61 (2.1)      4.31    0.08
Beckman Synchron CX3-7D, CX9ALX      15 (0.5)      4.05    0.09
Beckman Synchron CX4/5CE, 7/RTS      22 (0.8)      4.07    0.11
Beckman Synchron LX20                28 (1.0)      4.10    0.13
Dade Behring DIM/AR/MULT             35 (1.2)      4.10    0.10
Olympus 400-640/2700/5400            257 (8.9)     4.44    0.10
Olympus AU 5200                      37 (1.3)      4.24    0.09
Roche Cobas FARA/MIRA                14 (0.5)      4.34    0.21
Roche Cobas Integra                  125 (4.3)     4.23    0.11
Roche Modular                        233 (8.0)     4.25    0.11
Roche/Hitachi 717                    44 (1.5)      4.28    0.12
Roche/Hitachi 747                    74 (2.5)      4.22    0.14
Roche/Hitachi 911                    80 (2.8)      4.25    0.10
Roche/Hitachi 912                    65 (2.2)      4.22    0.11
Roche/Hitachi 917                    154 (5.3)     4.26    0.09
Toshiba TBA-FR series                22 (0.8)      4.26    0.15
VITROS 250 Chem System              573 (19.7)     3.81    0.15
VITROS 400, 700 Chem System          66 (2.3)      3.83    0.15
VITROS 500 550 Chem System           29 (1.0)      3.75    0.15
VITROS 950 Chem System              638 (22.0)     3.83    0.14
Roche Cobas Integra (BCG with RA)    239 (8.2)     4.23    0.10
All VITROS results                  1306 (45.0)    3.82    0.15
All BCG results except VITROS       1596 (55.0)    4.27    0.14
All BCG results                    2902 (100.0)    4.07    0.27

                                                   Albumin Results by
                                                   BCP, (a) g/dL

                                   Laboratories,      C-02
         Instrument (b)               No. (%)      Mean     SD

Designated comparison                              3.74
 method assigned value
Abbott Aeroset                       30 (1.1)      4.21    0.05
Beckman Synchron CX3-7D, CX9ALX     405 (14.9)     4.03    0.09
Beckman Synchron CX4/5CE, 7/RTS      171 (6.3)     4.03    0.09
Beckman Synchron CX4/5               18 (0.7)      4.04    0.12
Beckman Synchron LX20               617 (22.6)     3.99    0.08
Dade Behring DIM/AR/MULT            1407 (51.6)    4.09    0.08
Roche Hitachi 747                    12 (0.4)      4.11    0.11
Roche Hitachi 917                    30 (1.1)      4.13    0.10
All BCP results                     2726 (100)     4.06    0.09

                                                   Albumin Results
                                                   by BCG, (a) g/Dl

                                   Laboratories,    C-02
         Instrument (b)               No. (%)        SEM

Designated comparison
  method assigned value
Bayer ADVIA 1650                     61 (2.1)       0.010
Beckman Synchron CX3-7D, CX9ALX      15 (0.5)       0.023
Beckman Synchron CX4/5CE, 7/RTS      22 (0.8)       0.023
Beckman Synchron LX20                28 (1.0)       0.025
Dade Behring DIM/AR/MULT             35 (1.2)       0.017
Olympus 400-640/2700/5400            257 (8.9)      0.006
Olympus AU 5200                      37 (1.3)       0.015
Roche Cobas FARA/MIRA                14 (0.5)       0.056
Roche Cobas Integra                  125 (4.3)      0.010
Roche Modular                        233 (8.0)      0.007
Roche/Hitachi 717                    44 (1.5)       0.018
Roche/Hitachi 747                    74 (2.5)       0.016
Roche/Hitachi 911                    80 (2.8)       0.011
Roche/Hitachi 912                    65 (2.2)       0.014
Roche/Hitachi 917                    154 (5.3)      0.007
Toshiba TBA-FR series                22 (0.8)       0.032
VITROS 250 Chem System              573 (19.7)      0.006
VITROS 400, 700 Chem System          66 (2.3)       0.018
VITROS 500 550 Chem System           29 (1.0)       0.028
VITROS 950 Chem System              638 (22.0)      0.006
Roche Cobas Integra (BCG with RA)    239 (8.2)      0.006
All VITROS results                  1306 (45.0)     0.004
All BCG results except VITROS       1596 (55.0)     0.004
All BCG results                    2902 (100.0)     0.005

                                                    Albumin Results
                                                    by BCG, (a) g/Dl

                                   Laboratories,    C-02
        Instrument (b)               No. (%)        SEM

Designated comparison
 method assigned value
Abbott Aeroset                       30 (1.1)       0.009
Beckman Synchron CX3-7D, CX9ALX     405 (14.9)      0.004
Beckman Synchron CX4/5CE, 7/RTS      171 (6.3)      0.007
Beckman Synchron CX4/5               18 (0.7)       0.028
Beckman Synchron LX20               617 (22.6)      0.003
Dade Behring DIM/AR/MULT            1407 (51.6)     0.002
Roche Hitachi 747                    12 (0.4)       0.032
Roche Hitachi 917                    30 (1.1)       0.018
All BCP results                     2726 (100)      0.002

                                                   Albumin Results
                                                   by BCG, (a) g/Dl

                                   Laboratories,    C-02

         Instrument (b)               No. (%)      Difference

Designated comparison
  method assigned value
Bayer ADVIA 1650                     61 (2.1)       0.19 (*)
Beckman Synchron CX3-7D, CX9ALX      15 (0.5)        -0.07
Beckman Synchron CX4/5CE, 7/RTS      22 (0.8)        -0.05
Beckman Synchron LX20                28 (1.0)        -0.02
Dade Behring DIM/AR/MULT             35 (1.2)        -0.02
Olympus 400-640/2700/5400            257 (8.9)      0.32 (*)
Olympus AU 5200                      37 (1.3)       0.12 (*)
Roche Cobas FARA/MIRA                14 (0.5)         0.22
Roche Cobas Integra                  125 (4.3)        0.11
Roche Modular                        233 (8.0)      0.13 (*)
Roche/Hitachi 717                    44 (1.5)       0.16 (*)
Roche/Hitachi 747                    74 (2.5)       0.10 (*)
Roche/Hitachi 911                    80 (2.8)       0.13 (*)
Roche/Hitachi 912                    65 (2.2)       0.10 (*)
Roche/Hitachi 917                    154 (5.3)      0.14 (*)
Toshiba TBA-FR series                22 (0.8)       0.14 (*)
VITROS 250 Chem System              573 (19.7)     -0.31 (*)
VITROS 400, 700 Chem System          66 (2.3)      -0.29 (*)
VITROS 500 550 Chem System           29 (1.0)      -0.37 (*)
VITROS 950 Chem System              638 (22.0)     -0.29 (*)
Roche Cobas Integra (BCG with RA)    239 (8.2)      0.11 (*)
All VITROS results                  1306 (45.0)    -0.30 (*)
All BCG results except VITROS       1596 (55.0)       0.15
All BCG results                    2902 (100.0)      -0.05

                                                  Albumin Results
                                                  by BCG, (a) g/Dl

                                   Laboratories,      C-02
         Instrument (b)               No. (%)      Difference

Designated comparison
 method assigned value
Abbott Aeroset                       30 (1.1)       0.47 (*)
Beckman Synchron CX3-7D, CX9ALX     405 (14.9)      0.29 (*)
Beckman Synchron CX4/5CE, 7/RTS      171 (6.3)      0.29 (*)
Beckman Synchron CX4/5               18 (0.7)       0.30 (*)
Beckman Synchron LX20               617 (22.6)      0.25 (*)
Dade Behring DIM/AR/MULT            1407 (51.6)     0.35 (*)
Roche Hitachi 747                    12 (0.4)       0.37 (*)
Roche Hitachi 917                    30 (1.1)       0.39 (*)
All BCP results                     2726 (100)      0.32 (*)

Abbreviations: Chem, chemistry; RA, rapid absorbance.

(a) Values with an asterisk (*) are statistically different
from the value of the designated comparison method at P <.001.

(b) Instrument companies: Abbott Diagnostics, Chicago, Illinois;
Bayer Diagnostics (Siemens), Tarrytown, New York; Beckman Coulter,
Brea, California; Dade Behring (Siemens), Tarrytown, New York;
Olympus (Beckman), Brea, California; Roche Diagnostics,
Indianapolis, Indiana; Schiapparelli, West Caldwell, New Jersey;
Toshiba Medical, Tokyo, Japan; VITROS (Ortho Clinical Diagnostics),
Rochester, New York.

Table 2. Mean Albumin Values, Standard Deviations, and Standard
Errors of the Mean for Albumin Results by Bromocresol Green (BCG) and
Bromocresol Purple (BCP)  Methods From the College of American
Pathologists Chemistry Survey (2003) by Selected Instrument Groups
and the Designated Comparison Method

                                       Albumin by BCG, (a) g/dL

                                           Specimen C-01
   Instrument       Laboratories,
   Group (b)           No. (%)          Mean        SD       SEM

Designated
 comparison
 method                   2             3.80       0.03     0.021
All methods           2902 (100)        3.88 *     0.14     0.003
Advia (c)               61 (2.1)        3.95 *     0.08     0.010
Cobas Integra (c)     364 (12.5)        3.89 *     0.10     0.005
Hitachi (c)           640 (22.1)        3.97 *     0.10     0.004
Olympus (c)           257 (8.9)         4.03 *     0.10     0.006
Synchron (c)           65 (2.2)         3.75 *     0.09     0.011
VITROS (c)           1306 (45.0)        3.80       0.12     0.003

                                         Albumin by BCG, (a) g/dL
                                              Specimen C-01
Instrument          Laboratories,
Group (b)              No. (%)
                                        Mean        SD       SEM
Designated
 comparison
 method                   1             3.49       ...      ...
All methods           2726 (100)        3.74       0.09     0.002
Aeroset (e)            30 (1.1)         3.89       0.05     0.009
Dimension (e)        1407 (51.6)        3.78       0.07     0.002
Hitachi (e)            42 (1.5)         3.82       0.09     0.014
Synchron (e)         1211 (44.4)        3.69       0.08     0.002

                                         Albumin by BCG, (a) g/dL

                                             Specimen C-02
   Instrument       Laboratories,
   Group (b)           No. (%)          Mean        SD       SEM

Designated
 comparison
 method                   2             4.12         0.03     0.021
All methods           2902 (100)        4.07 *       0.26     0.005
Advia (c)                61 (2.1)       4.31 *       0.08     0.010
Cobas Integra (c)       364 (12.5)      4.23 *       0.11     0.006
Hitachi (c)             640 (22.1)      4.25 *       0.11     0.004
Olympus (c)             257 (8.9)       4.44 *       0.12     0.007
Synchron (c)             65 (2.2)       4.08         0.11     0.014
VITROS (c)             1306 (45.0)      3.82 *       0.15     0.004

                                         Albumin by BCG, (a) g/dL

                                              Specimen C-02
Instrument          Laboratories,
Group (b)              No. (%)
                                        Mean        SD       SEM
Designated
 comparison
 method                   1             3.74       ...      ...
All methods           2726 (100)        4.06       0.09     0.002
Aeroset (e)            30 (1.1)         4.21       0.05     0.009
Dimension (e)        1407 (51.6)        4.09       0.08     0.002
Hitachi (e)            42 (1.5)         4.12       0.10     0.015
Synchron (e)         1211 (44.4)        4.01       0.08     0.002

                                         Albumin by BCG, (a) g/dL

                                              Specimen C-03
   Instrument       Laboratories,
   Group (b)           No. (%)          Mean        SD       SEM

Designated
 comparison
 method                   2             2.61       ...      ...
All methods           2902 (100)        2.52 *     0.23     0.004
Advia (c)                61 (2.1)       2.71 *     0.05     0.007
Cobas Integra (c)       364 (12.5)      2.64 *     0.07     0.004
Hitachi (c)             640 (22.1)      2.74 *     0.09     0.003
Olympus (c)             257 (8.9)       2.73 *     0.07     0.004
Synchron (c)             65 (2.2)       2.56 *     0.11     0.013
VITROS (c)             1306 (45.0)      2.29 *     0.08     0.002

                                          Albumin by BCP, (d) g/dL

                                               Specimen C-03
Instrument          Laboratories,
Group (b)              No. (%)
                                        Mean        SD       SEM
Designated
 comparison
 method                   1             2.34       ...      ...
All methods           2726 (100)        2.51       0.07     0.001
Aeroset (e)            30 (1.1)         2.60       0.04     0.008
Dimension (e)        1407 (51.6)        2.54       0.06     0.002
Hitachi (e)            42 (1.5)         2.60       0.08     0.012
Synchron (e)         1211 (44.4)        2.47       0.06     0.002

                                         Albumin by BCP, (d) g/dL

                                              Specimen C-04
   Instrument       Laboratories,
   Group (b)           No. (%)          Mean        SD        SEM

Designated
 comparison
 method                   2             2.98        ...       ...
All methods           2902 (100)        2.99 *      0.19      0.003
Advia (c)                61 (2.1)       3.13 *      0.06      0.008
Cobas Integra (c)       364 (12.5)      3.06 *      0.08      0.004
Hitachi (c)             640 (22.1)      3.18 *      0.09      0.004
Olympus (c)             257 (8.9)       3.18 *      0.07      0.004
Synchron (c)             65 (2.2)       2.96        0.11      0.014
VITROS (c)             1306 (45.0)      2.81 *      0.08      0.002

                                         Albumin by BCP, (d) g/dL

                                               Specimen C-04
Instrument          Laboratories,
Group (b)              No. (%)
                                        Mean        SD        SEM
Designated
 comparison
 method                   1             2.72       ...       ...
All methods           2726 (100)        2.92       0.08      0.001
Aeroset (e)            30 (1.1)         3.03       0.05      0.009
Dimension (e)        1407 (51.6)        2.95       0.06      0.002
Hitachi (e)            42 (1.5)         3.02       0.06      0.010
Synchron (e)         1211 (44.4)        2.88       0.06      0.002

                                          Albumin by BCP, (d) g/dL

                                               Specimen C-05
   Instrument       Laboratories,
   Group (b)           No. (%)          Mean        SD       SEM

Designated
 comparison
 method                   2             4.36        0.03     0.027
All methods           2902 (100)        4.52 *      0.14     0.003
Advia (c)                61 (2.1)       4.53 *      0.09     0.012
Cobas Integra (c)       364 (12.5)      4.49 *      0.11     0.006
Hitachi (c)             640 (22.1)      4.54 *      0.10     0.004
Olympus (c)             257 (8.9)       4.64 *      0.10     0.006
Synchron (c)             65 (2.2)       4.35        0.13     0.016
VITROS (c)             1306 (45.0)      4.51 *      0.14     0.004

                                         Albumin by BCP, (d) g/dL

                                              Specimen C-05
Instrument          Laboratories,
Group (b)              No. (%)
                                        Mean        SD       SEM
Designated
 comparison
 method                   1             3.98       ...      ...
All methods           2726 (100)        4.37       0.10     0.002
Aeroset (e)            30 (1.1)         4.49       0.04     0.008
Dimension (e)        1407 (51.6)        4.40       0.08     0.002
Hitachi (e)            42 (1.5)         4.45       0.11     0.017
Synchron (e)         1211 (44.4)        4.32       0.09     0.003

(a) Values with an asterisk (*) using the BCG method are
statistically different from the value of the designated comparison
method at P < .001.

(b) Instrument companies: ADVIA (Siemens), Tarrytown, New York;
Aeroset (Abbott), Chicago, Illinois; Cobas Integra (Roche),
Indianapolis, Dimension (Siemens), Tarrytown, New York; Hitachi
(Roche), Indianapolis, Indiana; Olympus (Beckman), Brea, California;
Synchron (Beckman), Brea, California;VITROS (Ortho Clinical
Diagnostics), Rochester, New York.

(c) The results from these 6 instrument groups represent 92.8% of all
 BCG results reported to the College of American Pathologists.

(d) All mean albumin values by the BCP method from the selected
instrument groups are significantly different from those of the
designated comparison method.

(e) The results from these 4 instrument groups represent 98.7% of
all BCP results reported to the College of American Pathologists.

Table 3. Statistical Data for All Peer Groups for All Methods Based
on the Biuret Reaction for Total Protein From Specimen C-02
(Fresh Frozen Serum) in the College of American Pathologists
(CAP) Chemistry Survey (2003)

                                                 Total Protein, (a)
                                                      g/dL
                                                    C-02

        Instrument (b)            Laboratories,     Mean
                                     No. (%)

Reference measurement procedure                     7.16
CAP all data                       5701 (100)       7.15
Abbott Aeroset                      110 (1.9)       7.31
Bayer ADVIA 1650                    65 (1.1)        7.19
Beckman Synchron CX3-7D, CX9ALX     440 (7.7)       7.15
Beckman Synchron CX4/5CE, 7/RTS     185 (3.2)       7.23
Beckman Synchron CX4/5              20 (0.4)        7.26
Beckman Synchron LX20              611 (10.7)       6.86
Dade Behring DIM/AR/MULT           1402 (24.6)      7.22
Olympus 400-640/2700/5400           269 (4.7)       7.17
Olympus AU 5200                     38 (0.7)        7.24
Roche Cobas FARA/MIRA               38 (0.7)        7.14
Roche Cobas Integra                 382 (6.7)       7.09
Roche Modular                       231 (4.1)       7.22
Roche/Hitachi 717                   45 (0.8)        7.24
Roche/Hitachi 747                   87 (1.5)        7.26
Roche/Hitachi 911                   88 (1.5)        7.30
Roche/Hitachi 912                   69 (1.2)        7.28
Roche/Hitachi 917                   181 (3.2)       7.28
Schiapparelli ACE                   15 (0.3)        7.15
Toshiba TBA-FR Series               23 (0.4)        7.18
VITROS 250 Chem System             585 (10.3)       7.16
VITROS 400, 700 Chem System         65 (1.1)        7.18
VITROS 500 550 Chem System          31 (0.5)        7.21
VITROS 950 Chem System             638 (11.2)       7.13

                                  Total Protein, (a) g/dL
                                           C-02
        Instrument (b)               SD          SEM

Reference measurement procedure
CAP all data                        0.19        0.003
Abbott Aeroset                      0.11        0.010
Bayer ADVIA 1650                    0.15        0.019
Beckman Synchron CX3-7D, CX9ALX     0.16        0.008
Beckman Synchron CX4/5CE, 7/RTS     0.18        0.013
Beckman Synchron CX4/5              0.18        0.040
Beckman Synchron LX20               0.14        0.006
Dade Behring DIM/AR/MULT            0.12        0.003
Olympus 400-640/2700/5400           0.17        0.010
Olympus AU 5200                     0.12        0.019
Roche Cobas FARA/MIRA               0.27        0.044
Roche Cobas Integra                 0.18        0.009
Roche Modular                       0.14        0.009
Roche/Hitachi 717                   0.18        0.027
Roche/Hitachi 747                   0.12        0.013
Roche/Hitachi 911                   0.14        0.015
Roche/Hitachi 912                   0.13        0.016
Roche/Hitachi 917                   0.13        0.010
Schiapparelli ACE                   0.16        0.041
Toshiba TBA-FR Series               0.11        0.023
VITROS 250 Chem System              0.16        0.007
VITROS 400, 700 Chem System         0.17        0.021
VITROS 500 550 Chem System          0.23        0.041
VITROS 950 Chem System              0.17        0.007

                                    Total Protein, (a) g/dL
                                    C-02

        Instrument (b)                Bias

Reference measurement procedure
CAP all data                          0.01
Abbott Aeroset                      0.15 (*)
Bayer ADVIA 1650                      0.03
Beckman Synchron CX3-7D, CX9ALX      -0.01
Beckman Synchron CX4/5CE, 7/RTS     0.07 (*)
Beckman Synchron CX4/5                0.10
Beckman Synchron LX20              -0.30 (*)
Dade Behring DIM/AR/MULT            0.06 (*)
Olympus 400-640/2700/5400             0.01
Olympus AU 5200                     0.08 (*)
Roche Cobas FARA/MIRA                -0.02
Roche Cobas Integra                -0.07 (*)
Roche Modular                       0.06 (*)
Roche/Hitachi 717                     0.08
Roche/Hitachi 747                   0.10 (*)
Roche/Hitachi 911                   0.14 (*)
Roche/Hitachi 912                   0.12 (*)
Roche/Hitachi 917                   0.12 (*)
Schiapparelli ACE                    -0.01
Toshiba TBA-FR Series                 0.02
VITROS 250 Chem System                0.00
VITROS 400, 700 Chem System           0.02
VITROS 500 550 Chem System            0.05
VITROS 950 Chem System               -0.03

Abbreviation: Chem, chemistry.

(a) Values with an asterisk (*) are statistically different from the
value of the reference measurement procedure at P < .001.

(b) Instrument companies: Abbott Diagnostics, Chicago, Illinois;
Bayer Diagnostics (Siemens), Tarrytown, New York; Beckman Coulter,
Brea, California; Dade Behring (Siemens), Tarrytown, New York;
Olympus (Beckman), Brea, California; Roche Diagnostics,
Indianapolis, Indiana; Schiapparelli, West Caldwell,
New Jersey; Toshiba Medical, Tokyo, Japan; VITROS
(Ortho Clinical Diagnostics), Rochester, New York.

Table 4. Mean Total Protein Values, Standard Deviations, and
Standard Errors of the Mean for Selected Instruments
and the Reference Measurement Procedure From the College
of American Pathologists Chemistry Survey (2003)

                                        Total Protein, (a) g/dL

 Instrument      Laboratories,         Specimen C-01       Specimen
                                                             C-02
 Group (b,c)        No. (%)        Mean       SD     SEM       Mean

Reference
 measurement
 procedure             4           6.50      0.03   0.015      7.16

All methods       5701 (100)       6.44      0.17   0.002      7.15
ADVIA              65 (1.1)      6.57 *      0.14   0.017      7.19
Aeroset            110 (1.9)     6.61 *      0.10   0.010    7.31 *
Cobas Integra      382 (6.7)       6.48      0.17   0.009    7.09 *
Dimension         1402 (25.0)      6.49      0.11   0.003    7.22 *
Hitachi           701 (12.3)     6.54 *      0.12   0.005    7.26 *
Olympus            307 (5.4)     6.44 *      0.16   0.009      7.18
Synchron          645 (11.3)     6.43 *      0.13   0.005      7.18
Synchron LX-20    611 (10.7)     6.19 *      0.13   0.005    6.86 *
VITROS            1319 (23.1)    6.41 *      0.16   0.004      7.15

                                      Total Protein, (a) g/dL

 Instrument      Laboratories,     Specimen C-02     Specimen C-03
 Group (b,c)        No. (%)        SD       SEM       Mean

Reference
 measurement
 procedure             4          0.02     0.010      4.39

All methods       5701 (100)      0.19     0.003      4.43
ADVIA              65 (1.1)       0.15     0.019    4.51 *
Aeroset            110 (1.9)      0.10     0.010    4.50 *
Cobas Integra      382 (6.7)      0.18     0.009    4.46 *
Dimension         1402 (25.0)     0.12     0.003    4.42 *
Hitachi           701 (12.3)      0.14     0.005    4.52 *
Olympus            307 (5.4)      0.17     0.009    4.43 *
Synchron          645 (11.3)      0.17     0.007      4.39
Synchron LX-20    611 (10.7)      0.14     0.006    4.24 *
VITROS            1319 (23.1)     0.17     0.005    4.50 *

                                        Total Protein, (a) g/dL

 Instrument      Laboratories,     Specimen C-03     Specimen C-04
 Group (b,c)        No. (%)        SD       SEM       Mean

Reference
 measurement
 procedure             4          0.01     0.004      5.07

All methods       5701 (100)      0.13     0.002      5.11
ADVIA              65 (1.1)       0.09     0.011    5.21 *
Aeroset            110 (1.9)      0.10     0.010    5.20 *
Cobas Integra      382 (6.7)      0.12     0.006    5.13 *
Dimension         1402 (25.0)     0.08     0.002    5.12 *
Hitachi           701 (12.3)      0.09     0.004    5.20 *
Olympus            307 (5.4)      0.12     0.007    5.10 *
Synchron          645 (11.3)      0.09     0.004      5.08
Synchron LX-20    611 (10.7)      0.08     0.003    4.89 *
VITROS            1319 (23.1)     0.13     0.004    5.17 *

                                      Total Protein, (a) g/dL

 Instrument      Laboratories,    Specimen C-04     Specimen C-05
 Group (b,c)        No. (%)        SD       SEM       Mean

Reference
 measurement
 procedure             4          0.01     0.005      7.53

All methods       5701 (100)      0.15     0.002      7.45
ADVIA              65 (1.1)       0.10     0.012      7.58
Aeroset            110 (1.9)      0.10     0.009    7.65 *
Cobas Integra      382 (6.7)      0.14     0.007    7.48 *
Dimension         1402 (25.0)     0.09     0.002      7.54
Hitachi           701 (12.3)      0.10     0.004      7.55
Olympus            307 (5.4)      0.13     0.008    7.43 *
Synchron          645 (11.3)      0.11     0.004    7.45 *
Synchron LX-20    611 (10.7)      0.09     0.004    7.19 *
VITROS            1319 (23.1)     0.15     0.004    7.38 *

                                 Total Protein, (a) g/dL

 Instrument      Laboratories,    Specimen C-05
 Group (b,c)        No. (%)        SD       SEM

Reference
 measurement
 procedure             4          0.03     0.015

All methods       5701 (100)      0.19     0.003
ADVIA              65 (1.1)       0.14     0.018
Aeroset            110 (1.9)      0.11     0.010
Cobas Integra      382 (6.7)      0.19     0.010
Dimension         1402 (25.0)     0.13     0.004
Hitachi           701 (12.3)      0.14     0.005
Olympus            307 (5.4)      0.16     0.009
Synchron          645 (11.3)      0.14     0.006
Synchron LX-20    611 (10.7)      0.14     0.006
VITROS            1319 (23.1)     0.18     0.005

(a) Values with an asterisk (*) are statistically different from
the value of the reference measurement procedure at P < .001.

(b) The results from these 9 instrument groups represent 97.2%
of all results reported to the College of American Pathologists.

(c) Instrument companies: ADVIA (Siemens), Tarrytown, New York;
Aeroset (Abbott), Chicago, Illinois; Cobas Integra (Roche),
Indianapolis, Indiana; Dimension (Siemens), Tarrytown, New York;
Hitachi (Roche), Indianapolis, Indiana; Olympus (Beckman), Brea,
California; Synchron (Beckman), Brea, California; VITROS (Ortho
Clinical Diagnostics), Rochester, New York.

Table 5. Commutability Assessment Between the Conventional
Specimens and the C-02 Fresh Frozen Specimens
From the College of American Pathologists Chemistry
Survey (2003) (a,b)

                     C-01               C-02

Instrument       [DELTA] Means,  2 SD,     [DELTA] Means,
                     g/dL         g/dL          g/dL

Albumin--BCG comparative method
 ADVIA               0.15         0.16          0.19
 Cobas Integra       0.09         0.22          0.11
 Hitachi             0.17         0.22          0.13
 Olympus             0.23         0.24          0.32
 Synchron            -0.05        0.22          -0.04
 VITROS              0.00         0.30          -0.30

Albumin--BCP comparative method
 Aeroset             0.40         0.10          0.47
 Dimension           0.29         0.16          0.35
 Hitachi             0.33         0.20          0.38
 Synchron            0.20         0.16          0.27

Total protein--reference method
 ADVIA               0.07         0.30          0.03
 Aeroset             0.11         0.20          0.15
 Cobas Integra       -0.02        0.36          -0.07
 Dimension           -0.01        0.24          0.06
 Hitachi             0.04         0.28          0.10
 Olympus             -0.06        0.34          0.02
 Synchron            -0.07        0.34          0.02
 Synchron LX         -0.31        0.28          -0.30
 VITROS              -0.09        0.34          -0.01

                       C-03                 C-04

Instrument        [DELTA] Means,       [DELTA] Means,
                       g/dL                 g/dL

Albumin--BCG comparative method
 ADVIA                 0.10                 0.15
 Cobas Integra         0.03                 0.08
 Hitachi               0.13                 0.20
 Olympus               0.12                 0.20
 Synchron              -0.05                -0.02
 VITROS                -0.32                -0.17

Albumin--BCP comparative method
 Aeroset             0.26 (*)             0.31 (*)
 Dimension             0.20                 0.23
 Hitachi               0.26                 0.30
 Synchron              0.13                 0.16

Total protein--reference method
 ADVIA                 0.12                 0.14
 Aeroset               0.11                 0.13
 Cobas Integra         0.07                 0.06
 Dimension             0.03                 0.05
 Hitachi               0.13                 0.13
 Olympus               0.04                 0.03
 Synchron              0.00                 0.01
 Synchron LX           -0.15                -0.18
 VITROS                0.11                 0.10

                       C-05

Instrument        [DELTA] Means,
                       g/dL

Albumin--BCG comparative method
 ADVIA                 0.17
 Cobas Integra         0.13
 Hitachi               0.18
 Olympus               0.28
 Synchron              -0.01
 VITROS             0.15 (*,c)

Albumin--BCP comparative method
 Aeroset               0.51
 Dimension             0.42
 Hitachi               0.47
 Synchron            0.44 (*)

Total protein--reference method
 ADVIA                 0.05
 Aeroset               0.12
 Cobas Integra         -0.05
 Dimension             0.01
 Hitachi               0.02
 Olympus               -0.10
 Synchron              -0.08
 Synchron LX           -0.34
 VITROS                -0.15

Abbreviations: BCG, bromocresol green; BCP, bromocresol purple.

(a) [DELTA]  Means are the peer-group mean minus the comparative
or reference method mean.

(b) An asterisk (*) indicates a noncommutable sample
based on the [DELTA] mean for a conventional specimen being more
than [+ or -]2 SD different than the [DELTA]
mean for the C-02 (fresh frozen serum) sample for an
instrument group.

(c) Instrument companies: ADVIA (Siemens), Tarrytown, New York;
Aeroset (Abbott Diagnostics), Chicago, Illinois;
Cobas Integra (Roche), Indianapolis, Indiana; Dimension
(Siemens), Tarrytown, New York; Hitachi
(Roche), Indianapolis, Indiana; Olympus (Beckman), Brea,
California; Synchron (Beckman), Brea, California;
VITROS (Ortho Clinical Diagnostics), Rochester,
New York.


----------

Please note: Illustration(s) are not available due to copyright restrictions.
COPYRIGHT 2013 College of American Pathologists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Lo, Stanley F.; Miller, W. Greg; Doumas, Basil T.
Publication:Archives of Pathology & Laboratory Medicine
Article Type:Survey
Geographic Code:1USA
Date:Jul 1, 2013
Words:8576
Previous Article:Prevalence of terminal duct lobular units and frequency of neoplastic involvement of the nipple in mastectomy.
Next Article:Accuracy and false-positive rate of the cytologic diagnosis of follicular cervicitis: observations from the College of American Pathologists Pap...
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters