Comparison of the results of enzyme-linked immunosorbent assay (ELISA) to mass-spectrometry based analytical methods for six unregulated contaminants in source water and finished drinking-water samples

Regulatory entities, such as the Minnesota Department of Health, monitor public water systems for conformance with Federal and State monitoring requirements and water-quality standards. Although some contaminants have Federal and (or) State regulations and guidance values, many contaminants, such as pesticides and pharmaceuticals, are unregulated in that only non-enforceable health-based guidance values have been assigned to them. Furthermore, because these contaminants are not regulated, commonly only limited resources are available to public water systems or regulatory entities to monitor them in drinking water. Focused screening efforts on contaminants that are frequently detected in the environment can provide information to help monitoring entities prioritize their sampling efforts.

Here we assess the use of enzyme-linked immunosorbent assay (ELISA) method, a rapid, inexpensive screening method, as an alternative to more expensive methods to analyze source and finished drinking-water samples collected from public water systems throughout Minnesota for three commonly detected pesticides (atrazine, imidacloprid, and pyrethroids) and three commonly detected pharmaceuticals (caffeine, carbamazepine, and sulfamethoxazole). The ELISA results were compared to results provided by more advanced mass-spectrometry analytical methods at the U.S. Geological Survey National Water Quality Laboratory (NWQL) and SGS AXYS Analytical Services Ltd. (AXYS).

Overall, these datasets are highly censored (>80 percent) and contain multiple reporting limits within and between laboratories. To discern agreement between paired contaminant group results (target contaminant plus immunologically similar contaminants) by ELISA and the advanced analytical methods at NWQL and AXYS, presence-absence agreement analysis was coupled with false negative and false positive analysis. Analysis of presence-absence agreement shows that ELISA has generally good agreement (77.9 to 100 percent) with both NWQL and AXYS for all unregulated contaminant groups. Imidicloprid, pyrethroids, and caffeine contaminant groups have relatively low false positivity rates (16, 6, and 5 percent, respectively) when analyzed by ELISA, which indicates the ELISA method, for these contaminant groups, could be experiencing low-level interference attributed to the detection of immunologically similar contaminants. Similarly, sulfamethoxazole has a low false positivity rate (0.8 percent), which indicates ELISA is likely not overestimating results for this contaminant group. Analyses for carbamazepine and sulfamethoxazole by ELISA resulted in low false negativity rates (1.6 and 0.8 percent, respectively), which indicates the ELISA method is likely not underestimating the results for this contaminant group. Conversely, the atrazine contaminant group has a high false negativity rate (84 percent), which indicates the method has a strong negative bias and that ELISA underestimates results for this contaminant. These qualitative results indicate that the ELISA method could potentially serve as a reliable and cost-effective screening method to help drinking water monitoring entities prioritize sampling efforts for analyzing carbamazepine and sulfamethoxazole in source and finished drinking-water samples collected from public water systems. At the same time, although ELISA did not prove to be a good screening method for atrazine, evaluation of ELISA results indicated that its use for screening imidacloprid, pyrethroids, and caffeine could be beneficial for water testing.