The identification of chemicals that have the potential to cause injury to the immune system is of considerable public health significance, as alterations in immune function can lead to increased incidence of hypersensitivity disorders, autoimmune or infectious diseases or neoplasias. Experimental animal data collected over the past 15 years using standardized testing panels have provided a database from which the sensitivity and predictability of a variety of tests commonly used for the screening of chemicals for immunotoxicity have been evaluated. These results have been used as guidelines for risk assessment in immunotoxicity and have been the basis for a number of regulatory activities. In April 1998 a Workshop was held at NIEHS to establish study design to determine the sensitivity and predictability of extended histopathology as an indicator of immunotoxicity as compared with the National Toxicology Program functional testing battery. Standardized slide sets were generated for 10 chemicals, which had previously been evaluated for their immunotoxicity using functional tests. The histological evaluation of the slides has been completed and the data have been formatted for incorporation into the Immunotoxicology database. Russell Helms and Dr. Chris Portier, a collaborator on the original risk assessment studies, have nearly completed the data analysis of the pathology scoring. The data indicate that for a majority of the outcomes, there was good agreement between the pathologists. However, even for variables with excellent agreement there was some evidence of inconsistency among pathologists. A direct comparison of the ratings for each pathologist indicates that, even where there was good agreement, certain individuals tended to be more conservative while others were more able to discern subtle changes than others. Additional analyses examined the consistency of a pathologists ratings in a single tissue by investigating the correlation among all the measures in the same tissue type. When data from all pathologists were combined, measures in each of the four tissues seemed highly correlated with the other measures from that tissue. However, when correlations were examined for each pathologist independently, only the thymus evaluations maintained the same high degree of correlation. For spleen, bone marrow and lymph node measures, the outcomes for each pathologist were less well correlated, raising concern about the consistency of a pathologists ratings across different measures in each of these tissues. The final analysis will compare the ability of the histologic evaluation and function tests to determine whether a compound is considered immunotoxic.
