|
LLNA Report Appendix P Page 1 of 7: Comparison of LLNA versus GPMT/BA and Human Data, by Chemical and Product Class Previous Appendix 1 2 3 4 5 6 7 Next Page Return to the NICEATM-ICCVAM Document Search Page The tabulated LLNA data provided in Appendix A was used to compare, by chemical and product class, the sensitivity, specificity, positive predictivity, negative predictivity, and accuracy of the:
The results of these analysis are presented in the accompanying four tables. Tables 1 - 4 are based on a comparison by chemical class; Tables 5 - 8 by product class. The accuracy of each comparison are presented graphically in Figures 1 through 4. Center staff member Bonnie Carson, M.S. Organic Chemistry, assigned the chemical classes based on subsituent groups when a graphic molecular structure was readily available or could be drawn based on the chemical name. Some chemical class assignments, such as potential Michael-reactive agent, were based on assignments by Ashby et al. (1995). Chemical classes selected for the Center's analysis were generally those that possessed electrophilic moieties. The sources for the product classes were Budavari (1996), Truett (1998) and Chemfinder (1997). A chemical or product may be present in more than one chemical or product class and not all chemicals listed could be placed in one of the classes used. A number of these class/product comparisons are of very limited value considering the small number of chemicals tested in common among the various assays, and especially in terms of human sensitization results. To increase the number of possible comparisons to human data, all guinea pig test data were considered and human patch test allergens were included in the analyses. Their inclusion was based on an assumption that the substance would not be in use in a commercial test kit if it did not test positive in at least some individuals. In making these comparisons, unpublished data (as indicated in the Appendix) were included. Although several chemical or product classes are clearly underrepresented in these analyses, the correlation between the LLNA and guinea pig tests appeared to be disparate, by chemical class, only for lactones and salts. However, when compared against human sensitization results, the LLNA and GPT appear to be equal in accuracy. Previous Appendix 1 2 3 4 5 6 7 Next Page Return to the NICEATM-ICCVAM Document Search Page |
