I've only used rats as well, unfortunately.
I was going to suggest something similar to what Luke said. Rather than performing a CPA experiment that will likely need multiple conditioning days (depending on how aversive the drug is), can you look at depression of a normally rewarding activity as a measure of aversiveness? This doesn't need to be ICSS, but it can be something as simple as sucrose preference (i.e., does the intervention cause an animal to drink less sugar water?). This is likely a measure of 'anhedonia', but I think you could argue that anhedonia is aversive.