I have done my share of griping about the deluge of overly broad intellectual-property claims pouring out of the U.S. Patent and Trademark Office. So I suppose I should cheer when the courts try to block this stream. But this spring, a Hoover Dam of a ruling-in a patent case involving broad ownership rights to the latest generation of painkillers-raises more painful questions about our patent system than it resolves.

The case pits the University of Rochester against the drugmakers Pfizer and Pharmacia. The University of Rochester claims a broad patent on the underlying research for the insanely lucrative painkiller medications called cox-2 inhibitors-including bestselling Celebrex, made by Pharmacia. Pfizer, which is in the process of buying out Pharmacia, claims that it developed the drug independently and therefore shouldn't have to pay royalties to the university.

The stakes could hardly be higher. Celebrex, the leader of the cox-2 crowd, is what the drug companies like to call a blockbuster-earning Pharmacia more than $1.5 billion annually since it came on the market in 1999, and raking in more than $3 billion last year. Also hanging directly on the verdict is an additional, and almost as large, revenue stream from another cox-2 blockbuster called Vioxx, manufactured by Merck.

In his surprising decision, Judge David G. Larimer of the federal district court in western New York invalidated the University of Rochester's patent on cox-2 inhibitors, ruling that the university patent is too broad and nonspecific to be considered a full-fledged invention. Larimer's decision hinges largely on the fact that the University of Rochester scientists never took what he calls the "last, critical step" of actually isolating a drug compound themselves.