Ortho-McNeil's Formulation Patent on Ultracet Declared Invalid

In a decision (pdf file) released April 17th, the U.S. District Court for the District of New Jersey granted summary judgment that certain claims of U.S. Patent No. RE39,221 are invalid as obvious. The '221 patent claims combinations of acetaminophen and tramadol, marketed by Ortho-McNeil (JNJ) as Ultracet. Barr (BRL) and Caraco (CPD) are challenging the patent.

Ortho-McNeil had previously asserted U.S. Patent No. 5,336,691 (the predecessor patent to RE39,221) against Barr and Kali Labs (PRX). In early 2007, the district court granted summary judgment that the asserted claims of the '691 patent were invalid as anticipated and obvious.

While the litigation over the '691 patent was pending, Ortho-McNeil surrendered the '691 patent and requested a reissue from the USPTO. When the '691 patent reissued in 2006 as the '221 patent, Ortho-McNeil filed suit against Kali, Barr, Par and Caraco. By that time, several of the defendants were already selling generic versions of Ultracet.

As part of the reissue, Ortho-McNeil obtained new claims that were less likely to be found invalid or not infringed. Therefore, it was unclear whether the generic companies would be able to knock out the claims of the '221 reissue patent. In July 2007, Par and Kali decided not to take that risk, and settled their litigation with Ortho-McNeil, agreeing to cease selling their generic products by November of last year. Barr and Caraco remained in the litigation.

The asserted claims of the '221 patent are directed to a formulation "consisting essentially of" a tramadol-to-acetaminophen ratio that ranges from about 1:5 to about 1:19. Ultracet has a ratio of 1:8.67.

This month's decision largely draws from the 2007 summary judgment that invalidated the claims of the '691 patent as obvious and anticipated. Both opinions rely on U.S. Patent No. 3,652,589 (pdf file), which discloses various embodiments of a four-agent tablet that includes acetaminophen and tramadol as two of the active ingredients. In particular, the '589 patent discloses formulations where the acetaminophen-to-tramadol ratio falls within the ratio range claimed in the '221 patent.

Nevertheless, the '589 patent uses "consisting of" language in its description of the pharmaceutical formulations. In an anticipation context, this "consisting of" language would prevent the '589 patent's four-agent formulations from reading on the two-agent formulations claimed in the '221 patent. But what effect should this linguistic choice have in the context of obviousness?

Ortho-McNeil argued that the "consisting of" language in the '221 patent would not have instructed a person of ordinary skill in the art to combine only acetaminophen and tramadol into a two-agent tablet. The court, however, was not persuaded, concluding that "[h]ere, a person of ordinary skill in the art would not be misled by the 'consisting of' language."

Essentially, Ortho-McNeil appears to have argued that the "consisting of" language is an instance of "teaching away" from the two-agent formulations claimed in the '221 patent. But is it? Does one of skill in the relevant art have knowledge of patent lingo, such as the distinction between "comprising" and "consisting of"? If not, then the district court may be correct. This does pose an interesting question, though. "Teaching away," after all, is evaluated from the perspective of one of skill in the relevant art, and not from the perspective of a patent lawyer having ordinary skill in the relevant art.

The court also cited another piece of prior art that disclosed a three-agent formulation that included acetaminophen and tramadol in ratios similar to those claimed in the '221 patent. No "teaching away" appeared in this second reference. Thus, the Federal Circuit could affirm the judgment without having to consider whether "consisting of," as recited in the '589 prior art patent, can constitute teaching away.