Stanley F Lo1, W Greg Miller, Basil T Doumas. 1. Reference Standards Laboratory, Department of Pathology, Children's Hospital of Wisconsin, Milwaukee, WI 53226, USA. slo@mcw.edu
Abstract
CONTEXT: Discrepant results for serum constituents were observed among peer groups in the College of American Pathologists Comprehensive Chemistry Survey. OBJECTIVES: To assess the performance of serum albumin and total protein measurement procedures and to evaluate the commutability of the conventional survey specimens. DESIGN: A fresh frozen, off-the-clot serum sample was included along with 4 conventional survey specimens. The fresh frozen, off-the-clot serum sample was prepared in a manner expected to confer commutability with native clinical samples. RESULTS: For the fresh frozen, off-the-clot serum sample, the mean values for 17 peer-groups were -0.07 to 0.32 g/dL from the bromocresol green albumin designated comparison method, whereas 4 VITROS (Ortho Clinical Diagnostics, Rochester, New York) peer groups differed by -0.29 to -0.37 g/dL (15 of 21 differences [71%] had P < .001). For bromocresol purple albumin methods, the mean differences from the designated comparison method from 8 peer groups were 0.25 to 0.47 g/dL (all had P < .001). For total protein methods, 23 peer group mean values were -0.07 to 0.15 g/dL from the reference measurement procedure (12 of 24 [50%] had P < .001). The Beckman (Fullerton, California) Synchron LX20 had a bias of -0.30 g/dL (P <.001). The commutability of the conventional specimens was acceptable for 23 of 24 bromocresol green method-material combinations (96%) and 13 of 16 bromocresol purple albumin method-material combinations (81%). All (100%) of the 36 method-material combinations had acceptable commutability for total protein. CONCLUSIONS: One (2.2%) of the instrument systems (Synchron) using bromocresol green and none (0%) of the instrument systems using bromocresol purple had satisfactory total-error performance for albumin measurement. Differences in results between bromocresol green and bromocresol purple methods precluded using common reference intervals for interpreting results for serum albumin. Eight of 9 instrument systems (86.5%) had satisfactory total-error performance for total protein measurement.
CONTEXT: Discrepant results for serum constituents were observed among peer groups in the College of American Pathologists Comprehensive Chemistry Survey. OBJECTIVES: To assess the performance of serum albumin and total protein measurement procedures and to evaluate the commutability of the conventional survey specimens. DESIGN: A fresh frozen, off-the-clot serum sample was included along with 4 conventional survey specimens. The fresh frozen, off-the-clot serum sample was prepared in a manner expected to confer commutability with native clinical samples. RESULTS: For the fresh frozen, off-the-clot serum sample, the mean values for 17 peer-groups were -0.07 to 0.32 g/dL from the bromocresol green albumin designated comparison method, whereas 4 VITROS (Ortho Clinical Diagnostics, Rochester, New York) peer groups differed by -0.29 to -0.37 g/dL (15 of 21 differences [71%] had P < .001). For bromocresol purple albumin methods, the mean differences from the designated comparison method from 8 peer groups were 0.25 to 0.47 g/dL (all had P < .001). For total protein methods, 23 peer group mean values were -0.07 to 0.15 g/dL from the reference measurement procedure (12 of 24 [50%] had P < .001). The Beckman (Fullerton, California) Synchron LX20 had a bias of -0.30 g/dL (P <.001). The commutability of the conventional specimens was acceptable for 23 of 24 bromocresol green method-material combinations (96%) and 13 of 16 bromocresol purple albumin method-material combinations (81%). All (100%) of the 36 method-material combinations had acceptable commutability for total protein. CONCLUSIONS: One (2.2%) of the instrument systems (Synchron) using bromocresol green and none (0%) of the instrument systems using bromocresol purple had satisfactory total-error performance for albumin measurement. Differences in results between bromocresol green and bromocresol purple methods precluded using common reference intervals for interpreting results for serum albumin. Eight of 9 instrument systems (86.5%) had satisfactory total-error performance for total protein measurement.