An approach to analytical validation and testing of body fluid assays for the automated clinical laboratory

Background: Biochemical analysis of body fluids may lend insight into the pathogenesis of disease. Most commercially-available clinical laboratory methods are intended for blood-derived specimens and/or urine and laboratories must characterize the analytical performance of these methods for other specimen types such as body fluids. The aim of this work is to demonstrate one approach for characterizing analytical performance of assays for clinical analysis of body fluids. Methods: Residual waste samples were obtained from clinically ordered testing. Validation studies were performed for 8 chemistry analytes on Roche Cobas 6000 c501 (Roche Diagnostics, Inc.) analyzers. Accuracy, precision, analytical measurable range, reportable range, analytic sensitivity, interferences, analyte stability were assessed. Laboratory workflow for body fluid handling was designed based on results obtained. Results: Sample matrix interferences were not observed for the body fluids tested and assay reportable ranges used for serum were validated for body fluids. Dilution of body fluid specimens with saline demonstrated non-linear recovery of some enzyme activities. Assay imprecision was comparable to the manufacturer's claims for serum. The serum index thresholds for interference from hemoglobin and lipemia were lower for body fluids compared to manufacturer's stated limits for serum. Pretreatment of body fluid samples with hyaluronidase caused a 28% false increase in lipase activity, and 13% increase in total protein concentration. Ambient temperature analyte stability in body fluids was ≤24 h for most analytes compared to manufacturer's stated stability of 7 days for serum. In contrast to serum, lactate dehydrogenase (LDH) was labile in frozen body fluid specimens. Conclusions: Validation of analytical methods for body fluid testing is a necessary exercise to exclude potential matrix effects and set pre-analytic specimen quality criteria.