probit analysis


I have dose-response data (measurements vs. dead/not-dead) from several different treatments. For them I performed probit analyses to get 50% and 95% dose values (e.g. 50% dead at this dose in a treatment).
But how to compare between different treatments, i.e. 50% dose from treatment 1 against 50% dose from treatment 2. Thus something like a posthoc test for probit analyses???? or is it better do do a separate analysis with ANOVA or regression, resp.
Thanks a lot for replies.