Performance assessment and validation of a paramagnetic particle-based enzyme-linked immunosorbent assay for...

A commercial magnetic particle-based enzyme-linked immunosorbent assay (ELISA) kit for the insecticide chlorpyrifos [O,O-diethyl-O-(3,5,6-trichloro-2-pyridyl) phosphorothioate] was evaluated for its specificity, precision, and accuracy, its susceptibility to matrix interferences in agricultural and environmental surface waters, and its comparability to a gas chromatographic/flame photometric (GC/FPD) method for the determination of organophosphorus pesticides in natural waters. Repeatability, reproducibility, and accuracy studies show that the kit satisfies current U.S. Environmental Protection Agency criteria for the assessment of analytical methods. Observable matrix effects were found to be present in all of the environmental test waters, with the slopes of calibration curves generated in each of the test matrices deviating from that of the control matrix by as much as 16%. Specificity studies indicate that the chlorpyrifos polyclonal antibody adequately differentiates the target compound from other structurally similar organophosphorus pesticides, with the exception of its methyl analogue. Cross-reactivity with chlorpyrifos-methyl was approximately 37%, while reactivity with diazinon, pyridaphenthion, diclofenthion, bromiphos-ethyl, bromiphos-methyl, pirimiphos-ethyl, and chlorpyrifos oxon ranged from 1.6 to 10.7%. Cross-reactivity with pirimiphos-methyl, 3,5,6-trichloro-2-pyridinol, diethyl phosphate, and diethyl thiophosphate was negligible (<1%). Validation of the paramagnetic particle ELISA format was accomplished using water samples from two monitoring studies that were collected, split, and analyzed directly by ELISA and by GC/FPD. Results of the two analytical methods were then compared using standard t tests, regression analysis, and differences against mean measurement (bias) plots. While the agreement between the two methods was determined to be satisfactory, ELISA exhibits consistent positive bias in environmental matrices. Several preanalysis mitigation steps were suggested that may help moderate bias, but additional study is recommended to explicate the exact factors responsible for its consistent overestimation of results.