Evaluation of commercially available rapid assays: a manufacturer's perspective.

In the recent paper of O’Connor et al. (1), four rapid methods for the detection of toxins of Clostridium difficile were evaluated. I have two main criticisms regarding the methodology of this study. First, for the different methods examined, the specimen treatment procedures were described as follows: for the culture procedure, “stool specimens were cultured on the day of receipt”; for the cytotoxicity evaluation, “a filtrate of each stool specimen was prepared and stored at 20°C”; and for immunological detection, “the stool specimen was frozen at 84°C until tested.” However, these three methods of sample preparation and treatment were compared with no obvious control. Second, in the procedure described for the use of the rapid assay for the detection of toxin A from Oxoid Ltd., the paper states that for all methods used in the evaluation, “no rejection criteria were applied in respect of interval from collection to receipt of specimens,” but the Oxoid product instruction leaflet clearly states the following: “samples should be tested within 3 h of collection. If samples cannot be tested within this time, they should be stored in a refrigerator at 2 to 8°C and tested within 72 h.” The lack of regard for interval between specimen collection and receipt, together with the fact that specimens were frozen at 84°C until tested, leads me to conclude that the product instructions were not followed correctly in the evaluation of the Oxoid Ltd. assay. The instructions for specimen collection and storage are given to help ensure that the clinical performance characteristics quoted in the leaflet are reliably met in the user’s laboratory. In part, the purpose of the development process pursued by responsible manufacturers of diagnostic kits is to design sample preparation procedures and test protocols that are not only convenient to the end-user but also robust. Ensuring that the correct result is obtained in the field is of paramount importance to us. The work of O’Connor et al. raises issues for referees who consider papers describing the performance of commercially available products. I am sure that most kit manufacturers would be happy to supply copies of product instruction leaflets to enable referees to compare the use of the products by the authors with the instructions given by the manufacturers. Given the above criticisms and the failure of the authors to confirm that there may have been toxin A B strains in the population, I cannot accept the conclusion that the Oxoid test is not suitable for the routine testing for evidence of C. difficileassociated diarrhea.