The assessment of the bactericidal activity of surface disinfectants. I. A comparison of three practical tests.

Three tests intended for the assessment of the bactericidal activity of surface disinfectants are compared. They are the French AFNOR test (AFNOR NF T 72-190: détermination de l'activité antibactérienne pour la décontamination des surfaces--méthode des portes-germes), the German DGHM test (Deutsche Gesellschaft für Hygiene und Mikrobiologie: Flächendesinfektionsversuch unter praxisnahen Bedingungen) and the Leuven test. From the theoretical analysis of the different elements of the testing techniques in which the three methods differ, it was deduced that it should be easier to pass the AFNOR test, since relatively more disinfectant solution is available to kill the same number of bacteria, whereas the Leuven test is the more severe, because the bacterial inoculum is larger and the germs are more protected by organic matter. The same differences are found in practice. This is demonstrated by the comparison of the results for 38 disinfectant preparations: the microbicidal effect obtained (log reduction factor, RF) is the lowest in the results by the Leuven test, and the best results are obtained in the DGHM test; this difference is found statistically significant in the Wilcoxon rank sum test for paired samples It could be that because of the different norm (the Leuven test for 30 min requires only an RF of 4.0, whereas the norm for the other tests is 5.0), the number of preparations passing (or failing) a test is the same, or that at least a preparation passing the more severe test passes automatically the easier tests; this is not always the case If the notion of passing or failing a test is omitted, and the ranks of the test results are compared in pairs, then, however, a very significant correlation is found among the test results, although to a lesser extent for the AFNOR test versus the Leuven tests in case of Pseudomonas aeruginosa. It can be concluded that the three testing techniques yield comparable results (correlation of the ranks), although a preparation passing a severe test does not automatically pass a less severe test. In any case there is a marked difference in the severity of the testing techniques and this is not neutralized by adjusting the required norm only.