Farré, Marinella, Martínez, Elena, Hernando, M.-D., Fernández-Alba, Amadeo, Fritz, Johann, Unruh, Eckehardt, Mihail, Otilia, Sakkas, Vasilis, Morbey, Ana, Albanis, Triantafyllos, Brito, Fatima, Hansen, Peter D., and Barceló, Damià
Abstract: An inter-laboratory comparison exercise was conducted under the European Union funded project entitled: Screening Methods for Water Data Information in Support of the Implementation of the Water Framework Directive (SWIFT-WFD) and coordinated by the Consejo Superior de Investigaciones Científicas (CSIC), in order to evaluate the reproducibility of different toxicity tests based on the bioluminescence inhibition of Vibrio fischeri, for the rapid water toxicity assessment. For the first time, this type of exercise has been organized in Europe, and using different tests based on the same principle. In this exercise, 10 laboratories from 8 countries (Austria, Cyprus, Germany, Greece, Italy, Portugal, Romania, and Spain) took place, and a total number of 360 samples were distributed. During the exercise, six series of six samples were analyzed along 5 months. Every batch of samples was composed by three real samples and three standard solutions. The real samples were: a raw influent and the effluent of a wastewater treatment plant (WWTP), and a sample from a first settlement of the WWTP spiked with a mixture of toxicant standards. A final number of 330 (91.7%) samples was analyzed, 3300 values in duplicate were collected, and the results for each sample were expressed as the 50% effective concentration (EC50) values calculated through five points of dilution inhibition curves, after 5 and 15min of incubation times. A statistical study was initiated using 660 results. The mean values, standard deviations (σ), variances (σ 2), and upper and lower warning limits (UWL and LWL) were obtained, using the EC50 values calculated with the result from the participating laboratories. The main objectives of this toxicity ring study were to evaluate the repeatability (r) and reproducibility (R) when different laboratories conduct the test, the influence of complex matrix samples, the variability between different tests based on the same principle, and to determine the rate at which participating laboratories successfully completed tests initiated. In this exercise, the 3.93% toxicity values were outliers according with the Z-score values and the Dixon test. The samples with the greater number of outliers were those with the smallest variability coefficient, corresponding to the greater and the smaller toxicity level. No relation was found through the cluster analysis, between the final results and the different commercial devices involved. Testing by multiple commercial devices did not appear to reduce the precision of the results, and the variability coefficient for the exercise was nearby to the average value for past editions carried out at national level, where the different participants used the same commercial device. Stability of samples was also followed during the exercise. While statistical significance differences were not found for the greater part of samples, for the sample from the WWTP influent, a significant decrease of the toxicity value was found along this study. Nevertheless, this was a type of sample with a high toxicity level during all the exercise. On the other hand, in order to obtain the chemical characterization of real samples, those were analyzed by chromatographic techniques, using different sequential solid phase extraction (SSPE) procedures, followed by liquid chromatography coupled with mass spectrometry (LC–MS), and gas chromatography–mass spectrometry (GC–MS). Good agreement was found between the chemical analysis results and the toxicity level of the samples. [Copyright &y& Elsevier]