RESUMO
An international round-robin study on the Ames fluctuation test [ISO 11350, 2012], a microplate version of the classic plate-incorporation method for the detection of mutagenicity in water, wastewater and chemicals was performed by 18 laboratories from seven countries. Such a round-robin study is a precondition for both the finalization of the ISO standardization process and a possible regulatory implementation in water legislation. The laboratories tested four water samples (spiked/nonspiked) and two chemical mixtures with and without supplementation of a S9-mix. Validity criteria (acceptable spontaneous and positive control-induced mutation counts) were fulfilled by 92-100%, depending on the test conditions. A two-step method for statistical evaluation of the test results is proposed and assessed in terms of specificity and sensitivity. The data were first subjected to powerful analysis of variance (ANOVA) after an arcsine-square-root transformation to detect significant differences between the test samples and the negative control (NC). A threshold (TH) value based on a pooled NC was then calculated to exclude false positive test results. Statistically, positive effects observed by the William's test were considered negative, if the mean of all replicates of a sample did not exceed the calculated TH. By making use of this approach, the overall test sensitivity was 100%, and the test specificity ranged from 80 to 100%.