Fecal electrolyte testing for evaluation of unexplained diarrhea: Validation of body fluid test accuracy in the absence of a reference method

Abstract Background Validation of tests performed on body fluids other than blood or urine can be challenging due to the lack of a reference method to confirm accuracy. The aim of this study was to evaluate alternate assessments of accuracy that laboratories can rely on to validate body fluid tests in the absence of a reference method using the example of sodium (Na + ), potassium (K + ), and magnesium (Mg 2 + ) testing in stool fluid. Methods Validations of fecal Na + , K + , and Mg 2 + were performed on the Roche cobas 6000 c501 (Roche Diagnostics) using residual stool specimens submitted for clinical testing. Spiked recovery, mixing studies, and serial dilutions were performed and % recovery of each analyte was calculated to assess accuracy. Results were confirmed by comparison to a reference method (ICP-OES, PerkinElmer). Results Mean recoveries for fecal electrolytes were Na + upon spiking = 92%, mixing = 104%, and dilution = 105%; K + upon spiking = 94%, mixing = 96%, and dilution = 100%; and Mg 2 + upon spiking = 93%, mixing = 98%, and dilution = 100%. When autoanalyzer results were compared to reference ICP-OES results, Na + had a slope = 0.94, intercept = 4.1, and R 2  = 0.99; K + had a slope = 0.99, intercept = 0.7, and R 2  = 0.99; and Mg 2 + had a slope = 0.91, intercept = − 4.6, and R 2  = 0.91. Calculated osmotic gap using both methods were highly correlated with slope = 0.95, intercept = 4.5, and R 2  = 0.97. Acid pretreatment increased magnesium recovery from a subset of clinical specimens. Conclusions A combination of mixing, spiking, and dilution recovery experiments are an acceptable surrogate for assessing accuracy in body fluid validations in the absence of a reference method.