Evaluation and validation of a commercial ELISA for diazinon in surface waters

The performance of a commercially available microtiter plate ELISA kit for the determination of diazinon was evaluated for sensitivity, selectivity, i ntra-assay repeatability, accuracy, and matrix effects in fortified distill ed water and filtered and unfiltered environmental surface water samples. R epeatability and reproducibility studies show that the kit satisfies curren t EPA criteria for the assessment of analytical methods. Mean recoveries fr om spiked samples averaged 80.3, 95.5, and 103.5% from distilled, unfiltere d surface, and filtered surface waters, respectively. The experimentally de termined method detection limit (MDL) for the commercial diazinon microtite r plate format (0.0159 mu g L-1) was comparable to the least detectable dos e (LDD) established by the manufacturer(0.022 mu g L-1). Specificity studie s indicate that the diazinon polyclonal antibody can readily distinguish th e target compound from other structurally similar organophosphorus analogue s, with the exception of diazoxon. Cross-reactivity with the oxon was appro ximately 29%, while reactivity with pirimiphos-methyl, pirimiphos-ethyl, an d chlorpyrifos-ethyl was negligible. A slight matrix effect was discovered to be present in both filtered and unfiltered environmental water matrixes, but its effect on the immunoassays is insignificant within experimental er ror. For validation of the microtiter plate ELISA format, environmental sur face and storm runoff water samples were collected, split, and analyzed dir ectly by ELISA and by liquid-liquid extraction followed by GC (California S tate Department of Food and Agriculture method EM 46.0). Results of the two analytical methods were then compared statistically. A close correlation w as found between methods for unspiked and untreated river water samples (r = 0.969) while a much less robust correlation was obtained for runoff water s (r = 0.728). Results from runoff waters exhibit a particularly high posit ive bias for the ELISA method relative to the GC method. Cross-reactivity o f diazoxon and probably other unidentified cross-reacting components may be responsible for the exaggerated account of the target analyte in surface a nd runoff waters. While excellent for screening purposes, further study is required to elucidate and quantify the factors responsible for the consiste nt overestimation of ELISA results before the kit can be employed routinely for regulatory compliance monitoring.