Novartis sued Teva on U.S. Patent No. 6,162,802. The patent includes method of treatment claims and formulation claims directed to a combination of amlodipine and benazepril (marketed as Lotrel).
Novartis essentially lost on the first prong of the four-part preliminary injunction test: the likelihood of success on the merits. The main issue dealt with whether Teva’s amlodipine-benazepril formulation infringes the claims of the ‘802 patent. Particularly, the parties disagreed over whether the patent requires a physical barrier between the two active ingredients. (Teva’s product separates the incompatible ingredients by excipient, rather than by a barrier.)
The patent specification, and the prosecution history, seem to require that the two ingredients be “physically separated” from each other. Nevertheless, the patent’s broadest claim omits this limitation (while most dependent claims contain it). Novartis argued that claim differentiation prevents the court from reading the limitation into the broadest claim. On the other hand, Teva argued that disavowals in the specification and prosecution history, require reading the limitation into all claims. The court held off on deciding the dispute, but agreed that Teva’s argument has substantial merit.
But Novartis asserted that this makes no substantial difference. Teva, according to Novartis, “physically separate[s]” the two ingredients by means of its excipient, and therefore, still infringes the patent. Teva argued that the ‘802 patent’s use of the modifier “physical” implies that the separation must occur through use of a physical barrier. Since Teva uses no physical barrier in its formulation, it does not infringe.
The court rejected both parties’ renderings of this claim term. Novartis probably construes it too broadly, especially since none of the patent’s embodiments disclose a formulation where the active ingredients were interspersed in a sea of excipient. Yet Teva’s proposed construction is probably too narrow, failing to include embodiments in which the ingredients were in physical contact with each other. Hence, neither party proffered a credible construction of the claims.
This result necessitated a denial of the PI. Teva raised a substantial issue regarding whether the “physical barrier” imitation should be read into all of the claims. And, Novartis failed to provide a persuasive claim construction that captures Teva’s formulation.
In the end, the denial of the PI appears to be a reasonable result. The court basically recognized that PIs are best granted in situations where the court can make a straightforward assessment of claim scope. When this is not possible, patentees had better come to court with reasonable interpretations of the claim language (and not just proffer constructions that are most favorable to their interests). Anyway, it looks like we’re headed for one fun Markman showdown.