P24: Maximising the value of repeated dose toxicity data

Abstract

One of the main reasons cited for preclinical attrition is drug induced toxicity. As repeated dose toxicity tests are low throughput, expensive and time-consuming there is a clear benefit to organizations in capturing the data from these studies in a structure-searchable format. The data is then available for more effective internal use and can be mined for structure-activity relationships allowing read-across and the development of models for predicting toxicity of new chemicals. To this end we have proposed a database schema for collating repeated dose toxicity data. The proposed schema includes key experimental conditions such as species, strain, dose, frequency of administration and route. Subtables are then used to capture the results of any analyses performed. Subtable records are displayed in rows at the bottom of the parent record and this provides the flexibility needed for capturing study data when they may be none, a few or a significant number of findings reported. For serum chemistry, haematology, urinalysis and organ weight results are recorded as either “Increase”, “Decrease” or “No change” to allow searching across species where the reference ranges for specific parameters may vary. For clinical signs, gross pathology and histopathology controlled vocabulary is being used to avoid differences in terminology, spelling and transcriptional errors. The data can be queried using any of the fields in the schema, combining structure, CAS number and chemical name with specific events indicative of liver damage eg increased ALP and fibrosis. Feedback is being sought to determine if the proposed schema will meet the needs of organisations wishing to interrogate the data for structure activity relationships. It is planned to populate the new schema using subchronic and chronic study data from the US FDA CDER and CFSAN.