Derek Lowe points out that there’s no way aspirin would be approved today if it were a new drug. In fact, it wouldn’t even make it out of the lab. From Aspirin: Not Approvable:
Aspirin causes gastric lesions in rats and dogs, which are the standard small and large animal models for drug toxicity. This side effect occurs at levels which would raise red flags for any new compound. What would a present-day research organization do about it? If we stipulate that they could determine that aspirin worked by inhibiting cyclooxegenase enzymes, they would surely try to break the vascular effects of the drug apart from its anti-inflammatory effects. They would try to find new compounds that selectively inhibited only one of the enzyme subtypes. They would, in other words, produce Vioxx, and Celebrex, and the other COX-2 inhibitors, and this is just why these drugs were developed.
(Hat tip to Alex Tabarrok.)