The Risk of a Preliminary Answer: Injunction Bonds and the Generic Launch

By Dennis Crouch

Federal procedure has long used money bonds to price the risk that preliminary relief turns out wrong. Federal Rule of Civil Procedure 65(c) requires a party seeking preliminary injunctive relief to post a security bond to guarantee that the enjoined party will receive compensation if the injunction proves wrongful.

A preliminary injunction is relief granted on a forecast of the merits, issued after abbreviated discovery and a compressed hearing.  This is extraordinary relief (removing products from the market before any trial or judgment on the merits).  The rule thus requires the movant to provide security “in an amount that the court considers proper to pay the costs and damages sustained by any party found to have been wrongfully enjoined or restrained.”

In Otsuka America Pharmaceutical, Inc. v. Hetero Labs Ltd., No. 2025-2016 (Fed. Cir. July 1, 2026), the Federal Circuit rejected a district court order that skipped the injunction bond altogether. Judge Williams (D. Del.) preliminarily enjoined Hetero from launching its FDA-approved generic version of Nuedexta, Otsuka’s dextromethorphan and quinidine combination for treating pseudobulbar affect, and then waived the Rule 65(c) security entirely, finding Hetero’s risk of financial harm “speculative at best” and expressing concern that a multi-million-dollar bond would create “a chilling effect on access to justice.”

Writing for the panel, Judge Bryson affirmed the injunction itself, including a challenged claim construction. Judge Dyk dissented from that construction and would have dissolved the injunction. On the bond, though, the panel was unanimous: Third Circuit law treats Rule 65(c) security as nearly mandatory, and no recognized exception reaches an order blocking a commercial drug launch. The court vacated the waiver and remanded for the district court to set an amount for the bond.

What the bond does. When a court preliminarily enjoins a defendant, it restrains conduct before anyone has finally determined that the conduct is unlawful. If the forecast proves wrong, the defendant will have been barred, sometimes for years, from doing something it had a legal right to do. The bond is the mechanism the system uses to make that party whole, and the Supreme Court has explained that the bond is not merely security for a damages claim that would exist anyway; it is the source of the claim. “A party injured by the issuance of an injunction later determined to be erroneous has no action for damages in the absence of a bond.” W.R. Grace & Co. v. Local Union 759, 461 U.S. 757 (1983), citing Russell v. Farley, 105 U.S. 433 (1882). Absent a bond, and short of a malicious prosecution theory, a wrongfully enjoined party simply absorbs its own losses.

The requirement also disciplines the injunction request itself, since a plaintiff forced to stand behind its forecast with real money may think harder before seeking interim relief on a thin record.

A party is "wrongfully enjoined" under Rule 65(c) when it turns out to have had the legal right all along to do what the injunction prohibited, something usually established by a final judgment in its favor. Recovery is not automatic. The enjoined party must prove actual damages caused by the restraint, and it proceeds against the bond through the summary mechanism of Rule 65.1 rather than a separate suit.

Regional law and the Third Circuit’s near-mandate. Because the bond requirement is not unique to patent law, the Federal Circuit applies regional circuit law, here the law of the Third Circuit. That choice carried the day. The Third Circuit sits at the strict end of a genuine spread among the circuits, describing waiver as “so rare that the requirement is almost mandatory.” Frank’s GMC Truck Center, Inc. v. General Motors Corp., 847 F.2d 100 (3d Cir. 1988). The court has stated that it has “never excused” a district court from requiring a bond “where an injunction prevents commercial, money-making activities.” Zambelli Fireworks Manufacturing Co. v. Wood, 592 F.3d 412 (3d Cir. 2010). Other circuits are considerably more forgiving, treating the amount of security, including zero, as a matter committed to district court discretion.

For pharmaceutical practice, the geography matters as much as the doctrine. The District of Delaware and the District of New Jersey host the bulk of Hatch-Waxman litigation, and both sit within the Third Circuit. As a practical matter, then, the near-mandatory rule governs the paradigm generic-launch injunction. The stakes are typically large.  When Apotex launched generic Plavix at risk in 2006, the district court enjoined further sales and set security at $400 million (Apotex had asked for $4 billion). Sanofi-Synthelabo v. Apotex, Inc., 470 F.3d 1368 (Fed. Cir. 2006). In some ways, the bond fight is a preview of the damages case, with the amount set as the likely harm caused by wrongly enjoining generic sales through the end of trial.

The Federal Circuit’s response to the waiver was short. Hetero’s attempt to enter the market with a generic drug “is clearly a commercial, money-making activity,” which under Zambelli ends the inquiry.

The movant rarely deposits cash with the court. Instead, it buys the bond from a commercial surety, typically one of the large carriers such as Travelers, Liberty Mutual, or Chubb, paying an annual premium that runs a small percentage of the face amount, with large bonds often collateralized. The surety files a written undertaking with the clerk, a promise to pay up to the bond amount. If the defendant is later held wrongfully enjoined, it enforces that promise by motion under Rule 65.1, and the surety pays the defendant directly. The plaintiff must then reimburse the surety under its indemnity agreement, so the surety absorbs the loss only if the plaintiff cannot pay.

The salt-weight dispute, briefly. The merits fight deserves a quick discussion because of the split panel. Everyone agrees that both Nuedexta and the generic administer the drugs as salts, with each capsule containing 20 mg of dextromethorphan hydrobromide and 10 mg of quinidine sulfate. Measured by the salt weights actually administered, the ratio is exactly 1:0.5 and covered by the asserted claim. Convert to the active moieties, as Hetero urged, and the same capsule contains 15.4 mg of dextromethorphan and 8.7 mg of quinidine, a ratio of 1:0.56, outside the claim. The construction was therefore all-or-nothing: under Otsuka’s reading both products practice claim 1, and under Hetero’s reading neither does, Nuedexta included.

Judge Bryson found Otsuka’s reading better supported by the intrinsic record. Dependent claims 7 and 8 expressly cover administration of the compounds in pharmaceutically acceptable salt form, which means the independent claim’s terms must reach salts as well; the specification’s clinical studies use “DM” and “Q” as shorthand for salt quantities throughout; and the examiner characterized the pending claims as covering “dextromethorphan or dextromethorphan hydrobromide.” The majority also found it “highly improbable” that the patentee drafted claims excluding the very product the patent was meant to protect. Osram GmbH v. International Trade Commission, 505 F.3d 1351 (Fed. Cir. 2007).

Judge Dyk saw a construction unmoored from the invention. The point of the patent is minimizing quinidine in the body relative to dextromethorphan, and the counterion in a salt contributes nothing therapeutically; Hetero’s unrebutted expert described the salt forms as “merely formulation tools” for delivering the active compounds. He noted that the ratio limitation was not in the original claims at all but was added mid-prosecution to overcome an obviousness rejection, drawing on data generated after the priority date. On that record, Judge Dyk noted that a patentee who narrows its way past the prior art can easily narrow its way past its own product.