Inter and Intralaboratory Reproducibility of In Vitro Toxicological Assays

• Activity Number: 277
• Type: Contributed
• Date/Time: Tuesday, August 9, 2005 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Christina Bromley*+ and Nicola Richardson-Harman and Brigitte Beer and James Cummins and Carol Lackman-Smith and Ron Bromley and Patricia Reichelderfer
• Companies: BioStat Solutions, Inc. and BioStat Solutions, Inc. and Southern Research Institute and Southern Research Institute and Southern Research Institute and BioStat Solutions, Inc. and National Institutes of Health
• Keywords: Inter and intra-assay reproducibility ; cytoxicity assays ; multi-site comparison ; toxicity ; General Linear Model ; step-wise regression

Abstract

In vitro cytotoxicity evaluations are performed in preclinical testing to determine the safety of potential therapeutic agents. Kits, cell matrices, and methods may vary, reducing reproducibility of results. As part of a multisite comparison of N-9 (Nonoxynol-9, a spermicidal microbicide), we evaluated the reproducibility of historical N-9 data (1998--2004) from 121 toxicity assays. The viability of N-9-exposed samples was expressed relative to the cell control (100%) to compare assays with various toxicity endpoints. Standard Deviations (Log10) of toxicity levels from the same dose of N-9 were compared for 2-, 3-, and 5-fold intra-assay differences. The presence of a replicate effect indicated low interassay reproducibility (PROC GLM). An ANCOVA (N-9 covariate) and stepwise regression were used to test relative importance of assay variables. Curve-fit and point-point regression TC50 values were tested for their comparability. The GLM model produced the best fit (R2:0.81-0.99). ANCOVA increased sensitivity of reliability measures. Replicate as an independent variable successfully identified replicate effects. Intra-assay reproducibility was consistent at the two-fold level.