In Re Lipitor Antitrust Litigation

855 F.3d 126, 2017 WL 1359474

• Court: Court of Appeals for the Third Circuit
• Decided: April 13, 2017
• Docket: 14-4202, 14-4203, 14-4204, 14-4205, 14-4206, 14-4602 & 14-4632; 15-1184, 15-1185, 15-1186, 15-1187, 15-1274, 15-1323 & 15-1342
• Status: Published

Opinion

OPINION OF THE COURT

FISHER, Circuit Judge.

A pharmaceutical company holding the patent on a drug sues the manufacturer of a generic version of that drug for patent infringement. The patent-holder and the generic manufacturer later settle, with the former paying the latter not to produce a generic until the patents at issue expire. In FTC v. Actavis, Inc., — U.S. -, 133 S.Ct. 2223, 186 L.Ed.2d 343 (2013), the Supreme Court recognized that such a settlement — commonly known as a “reverse payment” — where large and unjustified, can sometimes unreasonably diminish competition in violation of the antitrust laws. To answer the antitrust question, Actavis explained, “it is not normally necessary to litigate patent validity” because “the size of the unexplained reverse payment can provide a workable surrogate for a patent’s weakness.” Id. at 2236-37.

These two sets of consolidated appeals involve allegations that the companies holding the patents for Lipitor and Effexor XR delayed entry into the market of generic versions of those drugs. The companies did so, plaintiffs say, by engaging in an overarching monopolistic scheme that involved fraudulently procuring and enforcing the underlying patents and then entering into a reverse-payment settlement agreement with a generic manufacturer. With a single exception, every complaint asserts one of these monopolization claims against the patent-holders. The cases were assigned to the same district judge, who ultimately dismissed the bulk of plaintiffs’ claims.

In this opinion, we address two questions of federal jurisdiction. First, do plaintiffs’ allegations of fraudulent procurement and enforcement of the patents require us to transfer these appeals to the Court of Appeals for the Federal Circuit? That court has exclusive jurisdiction over appeals from civil actions “arising under” patent law. 28 U.S.C. § 1295(a)(1). But not all cases presenting questions of patent law necessarily arise under patent law. See Christianson v. Colt Indus. Operating Corp., 486 U.S. 800, 108 S.Ct. 2166, 100 L.Ed.2d 811 (1988). Where, as here, patent law neither creates plaintiffs’ cause of action nor is a necessary element to any of plaintiffs’ well-pleaded claims, jurisdiction lies in this Court, not the Federal Circuit.

The second jurisdictional question we confront is confined to one of the Lipitor appeals, RP Healthcare, Inc. v. Pfizer, Inc., No. 14-4632. That case, brought by a group of California pharmacists, involves claims solely under California law and was filed in California state court. Following removal the District Court declined to remand the case to state court, citing potential patent defenses. That was error, as federal jurisdiction depends on the content of the plaintiffs complaint, not a defendant’s possible defenses. Before final judgment, however, the remaining non-diverse defendants were voluntarily dismissed, thus raising the possibility that, notwithstanding the District Court’s failure to remand the case, it possessed diversity jurisdiction before the time it- entered judgment. See Caterpillar Inc. v. Lewis, 519 U.S. 61, 117 S.Ct. 467, 136 L.Ed.2d 437 (1996). But because the state of the record before us is unclear with regard to the citizenship of the parties, we cannot reach the merits of this appeal until that question is resolved. We will accordingly remand the RP Healthcare appeal to the District Court so it can conduct jurisdictional discovery and address the matter in the first instance.

I

It is necessary to begin by discussing the regulatory framework that forms the foundation for the issues presented by these appeals.

A

“Apparently most if not all reverse payment settlement agreements arise in the context of pharmaceutical drug regulation, and specifically in the context of suits brought under statutory provisions allowing a generic drug manufacturer (seeking speedy marketing approval) to challenge the validity of a patent owned by an already-approved brand-name drug owner.” Actavis, 133 S.Ct. at 2227. With the Drug Price Competition and Patent Term Restoration Act, 98 Stat. 1585, as amended, known as the Hatch-Waxman Act, Congress “attempted to balance the goal of ‘makfing] available more low cost generic drugs’ with the value of patent monopolies in incentivizing beneficial pharmaceutical advancement.” King Drug Co. v. Smith-Kline Beecham Corp., 791 F.3d 388, 394 (3d Cir. 2015) (alteration in original) (quoting H.R. Rep. No. 98-857, pt. 1, at 14-15 .(1984)), cert. denied, — U.S. -, 137 S.Ct. 446, 196 L.Ed.2d 328 (2016). “The Act seeks to accomplish this purpose, in part, by encouraging ‘manufacturers of generic drugs ..to challenge weak or inval id patents on brand name drugs so consumers can enjoy lower drug prices.’ ” Id. (alteration in original) (quoting S. Rep. No. 107-167, at 4 (2002)). In Actavis, the Supreme Court identified four relevant features of Hatch-Waxman’s regulatory framework. 133 S.Ct. at 2227-29; see also King Drug, 791 F.3d at 394-96.

First, a drug manufacturer seeking to market a new, “pioneer” prescription drug must obtain approval from the Food and Drug Administration (FDA). See 21 U.S.C. § 356(b)(1). This approval process involves testing that is “long, costly, and comprehensive.” Actavis, 133 S.Ct. at 2228.

Second, following FDA approval of a brand-name drug, a generic manufacturer can file an Abbreviated New Drug Application (ANDA) indicating that the generic “has the same active ingredients as, and is biologically equivalent to, the brand-name drug.” Caraco Pharm. Labs., Ltd. v. Novo Nordisk A/S, 566 U.S. 399, 405, 132 S.Ct. 1670, 182 L.Ed.2d 678 (2012) (citing 21 U.S.C. § 355(j)(2)(A)(iv)). The ANDA process furthers drug competition “by allowing the generic to piggy-back on the pioneer’s approval efforts.” Actavis, 133 S.Ct. at 2228.

Third, the Hatch-Waxman Act “sets forth special procedures for identifying, and resolving, related patent disputes.” Id. The new drug applicant is required- to list any patents issued relating to the drug’s composition or methods of use. See 21 U.S.C. § 355(b)(1). If the FDA approves the new drug, it publishes this patent information, without verification, in its Orange Book (officially known as Approved Drug Products with Therapeutic Equivalence Applications). King Drug, 791 F.3d at 395 & n.5 (citing Caraco, 566 U.S. at 405-06, 132 S.Ct. 1670). In its ANDA, the generic manufacturer must “assure the FDA that its proposed generic drug will not infringe the brand’s patents.” Caraco, 566 U.S. at 406, 132 S.Ct. 1670. One method of assurance is known as “paragraph IV certification,” whereby the generic may assert that the relevant listed patents are “invalid or will not be infringed by the manufacture, use, or sale of the [generic] drug.” 21 U.S.C. § 355(j)(2)(A)(vii)(IV). The filing of a paragraph IV certification “means provoking litigation,” Caraco, 566 U.S. at 407, 132 S.Ct. 1670, as the patent statute treats it as an act of automatic infringement, see 35 U.S.C. § 271(e)(2)(A). If the brand-name patentee brings an infringement suit within 45 days, the FDA is required to withhold approving the generic for a 30-month period. If the courts decide the matter during that period, the FDA will follow that determination; if not, the FDA may move forward on its own. See 21 U.S.C. § 355(j)(5)(B)(iii).

Fourth, “Hatch-Waxman provides a special incentive for a generic to be the first to file an [ANDA] taking the paragraph IV route.” Actavis, 133 S.Ct. at 2228-29. From the time it begins marketing its generic, the first-filer enjoys a 180-day exclusivity period during which no other generic can compete with the brand-name drug. See 21 U.S.C. § 355(j)(5)(B)(iv). This exclusivity period “can prove valuable, possibly ‘worth several hundred million dollars.’ ” Actavis, 133 S.Ct. at 2229 (quoting C. Scott Hemphill, Paying for Delay: Pharmaceutical Patent Settlement as a Regulatory Design Problem, 81 N.Y.U.L. Rev. 1553, 1579 (2006)). The right to exclusivity belongs to the first-filer alone and is nontransferable. See 21 U.S.C. § 355(j)(5)(D). However, Hatch-Waxman does not preclude the underlying patent-holder from marketing a brand-generic version of its drug — known as an “authorized generic” — during the 180-day exclusivity period. See Mylan Pharm., Inc. v. FDA, 454 F.3d 270, 276-77 (4th Cir. 2006); Teva Pharm. Indus. Ltd. v. Craw ford, 410 F.3d 51, 55 (D.C. Cir. 2005); see also King Drug, 791 F.3d at 393; Sanofi-Aventis v. Apotex Inc., 659 F.3d 1171, 1174-75 (Fed. Cir. 2011).

B

In Actavis, the Supreme Court addressed whether reverse-payment settlements in the Hatch-Waxman context are subject to antitrust scrutiny. The Court concluded that such settlements “can sometimes violate the antitrust laws.” 133 S.Ct. at 2227. That is so, the Court held, because “[a]n unexplained large reverse payment itself would normally suggest that the patentee has serious doubts about the patent’s survival,” thus “suggesting] that the payment’s objective is to maintain supracompetitive prices to be shared among the patentee and the challenger rather than face what might have been a competitive market.” Id. at 2237.

Actavis rejected an approach known as the “scope of the patent” test, a near-categorical rule that “absent sham litigation or fraud in obtaining the patent, a reverse payment settlement is immune from antitrust attack so long as its anti-competitive effects fall within the scope of the exclusionary potential of the patent.” FTC v. Watson Pharm., Inc., 677 F.3d 1298, 1312 (11th Cir. 2012), rev’d sub nom. Actavis, — U.S. -, 133 S.Ct. 2223, 186 L.Ed.2d 343. The Court concluded that it would be “incongruous to determine antitrust legality by measuring the settlement’s anticompetitive effects solely against patent law policy, rather than by measuring them against procompetitive antitrust policies as well.” Actavis, 133 S.Ct. at 2231. Instead, the Court’s precedents “indicated that patent and antitrust policies are both relevant in determining the ‘scope of the patent monopoly’ — and consequently antitrust law immunity — that is conferred by a patent.” Id. The Court viewed these cases as “seekfing] to accommodate patent and antitrust policies, finding challenged terms and conditions unlawful unless patent law policy offsets the antitrust law policy strongly favoring competition.” Id. at 2233; see id. at 2244 (Roberts, C.J., dissenting) (“The majority seems to think that even if the patent is valid, a patent holder violates the antitrust laws merely because the settlement took away some chance that his patent would be declared invalid by a court.”). Finally, the Court observed, among other things, that “it is normally not necessary to litigate patent validity to answer the antitrust question (unless, perhaps, to determine whether the patent litigation is a sham).” Id. at 2236 (majority opinion). Such antitrust questions are to be addressed under the traditional rule-of-reason analysis. See id. at 2237-38.

II

In In re Lipitor Antitrust Litigation, Nos. 14-1402 et al., plaintiffs are a putative class of direct-purchasers of branded Lipitor, a putative class of end-payors, and several individual-retailers asserting direct-purchaser claims. We will refer to these three groups of plaintiffs collectively as the “Lipitor plaintiffs.” Defendants are Pfizer Inc., Ranbaxy Inc., and their respective corporate affiliates; they will be referred to collectively as the “Lipitor defendants.” There is also a fourth group of plaintiffs — California-based pharmacists raising claims under California law — that we will refer to independently as the “RP Healthcare plaintiffs.” In addition to suing the Lipitor defendants, the RP Healthcare plaintiffs also named additional parties as defendants whose relevance we will explore in Part V, infra.

l

Warner-Lambert Co. developed ator-vastatin, the active ingredient in its blockbuster brand-name drug Lipitor. One of the best-selling pharmaceutical products of all time, Lipitor reduces the level of bad LDL cholesterol in the bloodstream. Warner-Lambert, in partnership with Pfizer, launched Lipitor in 1997. The two companies merged in 2002, and we will refer to them collectively as “Pfizer.”

In 1987, Pfizer obtained the original patent for Lipitor. That patent — designated U.S. Patent No. 4,681,893 (the ‘893 Patent) — claims protection for atorvastatin. Initially scheduled to expire in May 2006, Pfizer eventually secured extensions on the ‘893 Patent’s term through March 24, 2010. Pfizer obtained additional, follow-on patent protection for Lipitor in December 1993, when the Patent and Trademark Office (PTO) issued U.S. Patent No. 5,273,-995 (the ‘995 Patent). That patent claims atorvastatin calcium, the specific salt form of the active atorvastatin molecule in Lipitor. The Lipitor plaintiffs assert that Pfizer committed fraud with regard to the procurement and enforcement of the ‘995 Patent. In particular, the Lipitor plaintiffs allege that Pfizer submitted false and misleading data to the PTO to support its claim that the cholesterol-synthesis inhibiting activity of atorvastatin calcium was surprising and unexpected. The ‘995 Patent expired on June 28, 2011. Following Lipitor’s 1997 launch, Pfizer obtained five additional patents, all of which, according to the Lipitor plaintiffs, could not block further generic versions of the drug from coming to market. Pfizer listed all Lipitor patents in the FDA’s Orange Book, with the exception of the process patents, which cannot be listed. The Lipitor plaintiffs allege fraud only with regard to the procurement and enforcement of the ‘995 Patent.

After obtaining ANDA first-filer status for generic Lipitor in August 2002, Ran-baxy notified Pfizer of its paragraph IV certifications, which contended that none of the valid patent claims that covered Lipitor would be infringed by the sale, marketing, or use of its generic. Pfizer sued Ranbaxy in the District Court for the District of Delaware within the 45-day period prescribed by Hatch-Waxman, alleging that Ranbaxy’s generic would infringe the ‘893 and ‘995 Patents. Pursuant to Hatch-Waxman, the filing of Pfizer’s lawsuit stayed FDA approval of Ranbaxy’s ANDA for 30 months.

After a bench trial, the district court ruled that Pfizer’s patents were valid and enforceable and would be infringed by Ranbaxy’s generic. Pfizer Inc. v. Ranbaxy Labs. Ltd., 405 F.Supp.2d 495, 525-26 (D. Del. 2005). On appeal, the Federal Circuit largely agreed, affirming the district court’s ruling that the ‘893 Patent would be infringed. Pfizer Inc. v. Ranbaxy Labs. Ltd., 457 F.3d 1284, 1286 (Fed. Cir. 2006). The Federal Circuit reversed in part, however, holding that claim 6 of the ‘995 Patent was invalid due to what amounted to a scrivener’s error in the drafting of the claim. Id. at 1291-92. On remand, the district court enjoined FDA approval of Ran-baxy’s ANDA until March 24, 2010, the date of the ‘893 Patent’s expiration. Also in response to the Federal Circuit’s ruling, Pfizer applied for a reissuance of the ‘995 Patent to cure the drafting error. Ranbaxy filed an objection to the reissuance with the PTO.

In July 2005, as the 30-month statutory window halting Ranbaxy’s generic market entry was closing, Pfizer filed a citizen petition with the FDA stating that the amorphous noncrystalline form of atorvas-tatin used in generic Lipitor (including Ranbaxy’s, as identified in its ANDA) may be “inferior in quality” to branded Lipi tor’s crystalline form. Lipitor J.A. 1851. The Lipitor plaintiffs claim that this citizen petition was a sham. In May 2006, the FDA informed Pfizer that it had not yet reached a decision, citing the need for further review and analysis. The FDA denied the petition in a 12-page decision issued on November 30, 2011.

Around the same time as their Lipitor patent dispute, Pfizer and Ranbaxy were also locked in patent-infringement litigation regarding a separate drug called Ac-cupril. After Ranbaxy received ANDA approval and began marketing a generic Accupril product in conjunction with Teva Pharmaceuticals, Pfizer sued Ranbaxy and Teva in the District of New Jersey. On March 25, 2005, the district court issued a preliminary injunction halting Ranbaxy’s sales of generic Accupril, subject to Pfizer posting a $200 million bond to cover Ranbaxy’s damages in the event the injunction was improvidently granted. The Federal Circuit affirmed without prejudice to an ultimate resolution of the merits. Pfizer Inc. v. Teva Pharm. USA, Inc., 429 F.3d 1364, 1383 (Fed. Cir. 2005). On June 13, 2007, in light of the disputed patent’s expiration, the district court vacated the preliminary injunction. The only issues that remained contested were Pfizer’s limited claims for past damages and Ranbaxy’s counterclaim as secured by the preliminary injunction bond.

In March 2008, Pfizer again sued Ran-baxy in the District of Delaware, this time claiming that Ranbaxy’s generic Lipitor would infringe Pfizer’s two Lipitor-related process patents. Not long after, on June 18, 2008, Pfizer and Ranbaxy publically announced that they had reached a near-global litigation settlement — which the Lipitor plaintiffs allege constituted an unlawful reverse payment — regarding scores of patent litigations around the world, including the Lipitor and Accupril disputes. In particular, the settlement ended the Accupril litigation with prejudice, all domestic patent infringement litigation between Pfizer and Ranbaxy pertaining to Lipitor, and all foreign litigation between the two companies over Lipitor. As a result of the settlement, Ranbaxy received a licensed entry date of November 30, 2011 for generic Lipitor, Pfizer and Ranbaxy negotiated similar market entry dates for generic Lipitor in several foreign jurisdictions, Ranbaxy paid $1 million to Pfizer in connection with the Accupril litigation, and Pfizer’s $200 million injunction bond from the Accupril litigation was cancelled. Ran-baxy also withdrew its objection to the ‘995 Patent’s reissuance. The PTO reissued the ‘995 Patent in March 2009.

As part of the agreement, Ranbaxy delayed entry of its generic to March 2010, when the ‘893 Patent was set to expire. Due to its ANDA first-filer status, Ran-baxy was entitled to 180 days of market exclusivity. The Pfizer-Ranbaxy agreement consequently had the effect of maintaining a bottleneck over the entry of generic Lipitor from later ANDA filers. Any other would-be generic manufacturer that wanted the 180-day period to begin earlier than November 2011 would need a court to hold that all of Pfizer’s Orange Book-listed patents were invalid or not infringed. Pfizer helped to forestall this possibility, the Lipitor plaintiffs say, through a combination of several lawsuits against subsequent ANDA filers. The FDA approved Ran-baxy’s Lipitor ANDA on November 30, 2011, the day Ranbaxy’s license to the unexpired Lipitor patents commenced.

Beginning in November 2011, the Lipitor direct-purchasers and end-payors, as well as the RP Healthcare plaintiffs, filed separate antitrust actions in various federal jurisdictions. The cases were referred to the Judicial Panel on Multidistrict Litigation (JPML) for coordination. In January 2012, the RP Healthcare plaintiffs withdrew their federal suit and refiled in California state court raising claims solely under California law. That suit was removed to federal court two months later.

The JPML transferred each case to the District of New Jersey, and assigned the matters to Judge Peter G. Sheridan. See In re Lipitor Antitrust Litig., 856 F.Supp.2d 1355 (J.P.M.L. 2012); In re Lipitor Antitrust Litig., 2012 WL 4069565 (J.P.M.L. Aug. 3, 2012). Thereafter, the direct-purchaser and end-payor plaintiffs filed amended class action complaints; the individual-retailer plaintiffs likewise filed complaints joining the consolidated proceedings. The complaints are substantively identical, raising the same two claims: First, a monopolization claim under section 2 of the Sherman Act (15 U.S.C. § 2) or a state analogue against Pfizer, asserting that the company engaged in an overarching anticompetitive scheme that involved fraudulently procuring the ‘995 Patent from the PTO (Walker Process fraud), enforcing the ‘995 Patent and certain process patents through sham litigation, filing a sham citizen petition with the FDA, and entering into a reverse-payment settlement with Ranbaxy. Second, the Lipitor plaintiffs raise a claim under section 1 of the Sherman Act (15 U.S.C. § 1) or a state analogue against both Pfizer and Ranbaxy, challenging the reverse-payment settlement as an unlawful restraint of trade. We will refer to these claims, respectively, as the “section 2 monopolization claim” and the “section 1 restraint of trade claim.”

The RP Healthcare plaintiffs’ amended complaint raises an altogether different claim under California’s antitrust statute, the Cartwright Act, Cal. Bus. & Prof. Code § 16700 et seq. They allege that Pfizer, Ranbaxy, a Japanese company called Daiichi Sankyo (and an affiliate), and two large pharmacies entered into a per se unlawful market.allocation agreement regarding Lipitor. This agreement, according to the RP Healthcare plaintiffs, extended the life of Pfizer’s Lipitor-related patents and fixed prices for Lipitor and its generic equivalents at supracompetitive levels.

The Lipitor defendants filed motions to dismiss all complaints under Federal Rule of Civil Procedure 12(b)(6). On October 19, 2012, the District Court denied the RP Healthcare plaintiffs’ motion to remand to California state court, reasoning that “there may be many patent issues raised as defenses in this case which would engender federal jurisdiction.” Lipitor J.A. 2. And on May 16, 2013, the District Court stayed proceedings pending the Supreme Court’s decision in Actavis. In light of Actavis, the District Court reopened the case and permitted the parties to file supplemental briefs on the pending motions to dismiss.

On September 5, 2013, the District Court dismissed the Lipitor plaintiffs’ complaints to the extent they were based on anything other than the reverse-payment settlement. In re Lipitor Antitrust Litig., 2013 WL 4780496 (D.N.J. Sept. 5, 2013). In particular, the District Court rejected the Walker Process, sham litigation, and sham FDA citizen petition aspects of the Lipitor plaintiffs’ monopolization claims. Id. at *15-23. The court also granted leave to file amended complaints focused solely on the Pfizer-Ranbaxy reverse payment. Id. at *25-27.

The Lipitor plaintiffs filed amended complaints in October 2013. The direct-purchasers and end-payors attached their prior complaints as exhibits to their new complaints to preserve for appeal the allegations that had been dismissed. For their part, the independent-retailers stated in the first paragraph of their new complaints that they were also preserving the previously dismissed claims.

In November 2013, the Lipitor defendants once again moved to dismiss. On September 12, 2014, the District Court dismissed with prejudice the Lipitor direct-purchasers’ remaining argument that the Pfizer-Ranbaxy settlement was unlawful under Actavis. In re Lipitor Antitrust Litig., 46 F.Supp.3d 523 (D.N.J. 2014). The complaints of the end-payor, individual-retailer, and RP Healthcare plaintiffs were subsequently dismissed with prejudice in light of the District Court’s opinion.

The direct-purchasers filed a motion to amend the judgment and for leave to file an amended complaint, arguing that the District Court applied a novel pleading standard. That motion was denied on March 17, 2015. Lipitor J.A. 151-52. These timely appeals followed.

In In re Effexor XR Antitrust Litigation, Nos. 15-1184 et al., plaintiffs are a putative class of direct-purchasers of branded Effexor XR, a putative class of end-payors, two individual third-party pay-ors, and several individual-retailers asserting direct-purchaser claims. We will refer to these parties collectively as the “Effexor plaintiffs.” Defendants are Wyeth, Inc., Teva Pharmaceutical Industries Ltd., and their respective corporate affiliates. We will likewise refer to these parties collectively as the “Effexor defendants.”

In 1985, the PTO issued a patent for the compound venlafaxine hydrochloride. That patent was assigned to American Home Products, Wyeth’s predecessor. Eight years later, in 1993, the FDA granted Wyeth approval to begin marketing Effe-xor, a drug used to treat major depression. Effexor’s active ingredient is venlafaxine hydrochloride; the patent for that compound expired on June 13, 2008. In 1997, Wyeth introduced Effexor XR, an extended release, once-daily version. Wyeth obtained three patents for Effexor XR, all of which expired on March 20, 2017. The Effexor plaintiffs contend that Wyeth obtained the Effexor XR patents through fraud on the PTO, improperly listed those patents in the FDA’s Orange Book, and enforced those patents through serial sham litigation.

On December 10, 2002, Teva filed a paragraph IV certification challenging the validity of Wyeth’s Effexor XR patents. As the first company to file an ANDA with a paragraph IV certification for generic Ef-fexor XR, Teva was entitled to Hatch-Waxman’s 180-day period of marketing exclusivity. Wyeth brought suit against Teva for patent infringem§nt in the District of New Jersey.

In October 2005, shortly after the district court held a Markman hearing on claim construction, Wyeth and Teva reached a settlement. Under the settlement, which the Effexor plaintiffs allege constitutes an unlawful reverse payment, Wyeth and Teva reached an agreed-upon entry date of July 1, 2010 for generic Effexor XR, nearly seven years before the expiration of Wyeth’s patents related to that drug. Wyeth further agreed that it would not market an authorized-generic Effexor XR during Teva’s 180-day exclusivity period. In return, Teva would pay Wyeth royalties for the license, beginning at 15% during the 180-day period. If Wyeth chose not to introduce an authorized-generic after 180 days and no other generic entered the market, Teva was required to pay Wyeth 50% royalties for the next 180 days and 65% thereafter for up to 80 months. Moreover, in accordance with the settlement, Wyeth granted Teva a li cense to begin selling generic immediate release Effexor (Effexor IR) for two years prior to the June 2008 expiration of the original venlafaxine hydrochloride patent and agreed that it would not compete with Teva’s marketing of generic Effexor IR during that two-year period. Teva, for its part, would pay Wyeth 28% royalties during the first year and 20% during the second year.

Wyeth and Teva filed the settlement agreement with the district court presiding over the patent infringement litigation. In accordance with a 2002 consent decree, the Federal Trade Commission (FTC) had the right to weigh in on Wyeth’s settlements and to raise objections in advance. It offered no objection. The settlement was also submitted to the FTC and the U.S. Department of Justice pursuant to section 1112 of the Medicare Prescription Drug, Improvement, and Modernization Act of 2003, Pub. L. No. 108-173, 117 Stat. 2066, 2461-63 (2003) (codified at 21 U.S.C. § 355 note). The district court thereafter entered orders vacating its prior Markman rulings, dismissing the case, and adopting the terms of the settlement as a consent decree and permanent injunction. Effexor J.A. 1298.

Following the Wyeth-Teva settlement, between April 2006 and August 2011, Wyeth brought patent infringement suits against sixteen other companies that sought to market a generic Effexor XR. All suits settled under terms stipulating that Wyeth’s patents were valid and infringed.

Beginning in May 2011, several direct-purchasers of Effexor XR filed class action complaints in the Southern District of Mississippi challenging the lawfulness of the Wyeth-Teva settlement agreement. The cases were consolidated and, on September 21, 2011, the court transferred the action to the District of New Jersey.

After transfer, the direct-purchasers filed an amended consolidated class action complaint, a group of end-payors joined the case with a consolidated class action complaint of their own, several individual-retailers filed complaints, and two individual third-party payors together filed their own complaint. The complaints are substantially similar: Each alleges a monopolization claim against Wyeth under section 2 of the Sherman Act or analogous state statutes, asserting that Wyeth fraudulently induced the PTO to issue the three patents covering Effexor XR (Walker Process fraud), wrongfully listed those patents in the Orange Book, enforced those patents through serial sham litigation, and entered into a reverse-payment settlement with Teva. The complaints also raise a claim under section 1 of the Sherman Act or a state analogue against both Wyeth and Teva, challenging the reverse-payment settlement as an unlawful restraint of trade. As with the Lipitor appeals, we will refer to these claims, respectively, as the “section 2 monopolization claim” and the “section 1 restraint of trade claim.” (Though otherwise similar to the other complaints, the individual third-party payors’ complaint names only Wyeth and its affiliates as defendants. They also raise additional claims not relevant to these appeals.)

The Effexor defendants filed motions to dismiss under Rule 12(b)(6), but the District Court stayed proceedings pending the Supreme Court’s decision in Actavis. After Actavis was issued, the District Court vacated the stay, reopened the case, and called for supplemental briefing on the pending motions to dismiss. On October 23, 2013, the direct-purchasers (but no other party) filed an amended complaint.

On October 6, 2014, the District Court granted in part and denied in part the Effexor defendants’ motions to dismiss. In re Effexor XR Antitrust Litig., 2014 WL 4988410 (D.N.J. Oct. 6, 2014). It rejected the Effexor plaintiffs’ challenges to the Wyeth-Teva reverse-payment settlement and dismissed with prejudice the section 1 restraint of trade claims. Id. at *19-24. However, the District Court declined to dismiss the Effexor plaintiffs’ Walker Process allegations against Wyeth. Id. at *24-26. At the Effexor plaintiffs’ request, the court granted final judgment on the restraint of trade claims under Federal Rule of Civil Procedure 54(b).

These timely appeals followed. On February 27, 2015, the Effexor defendants moved this Court to transfer the Effexor appeals to the Federal Circuit on the ground that the Effexor plaintiffs’ complaints assert claims that arise under patent law. We denied the motion without prejudice to the Effexor defendants raising the jurisdictional argument in their merits briefs.

Ill

The District Court possessed subject-matter jurisdiction, at a minimum, under the following statutes: With respect to the Lipitor and Effexor direct-purchasers and independent-retailers, the District Court had jurisdiction under 28 U.S.C. §§ 1331 and 1337(a). With respect to the Lipitor and Effexor end-payors, the District Court had jurisdiction under 28 U.S.C. § 1332(d). And with respect to the Effexor independent third-party payors, the District Court had jurisdiction under 28 U.S.C. § 1332(a)(1) and (3).

The Lipitor and Effexor defendants contend that the District Court also had jurisdiction over each of these cases under 28 U.S.C. § 1338(a), thus necessitating transfer of these appeals to the Federal Circuit. The RP Healthcare plaintiffs, for then-part, argue that the District Court did not possess subject-matter jurisdiction at all; they say their case properly belongs in California state court.

Though our jurisdiction to reach the merits of these appeals is disputed, “it is familiar law that a federal court always has jurisdiction to determine its own jurisdiction.” United States v. Ruiz, 536 U.S. 622, 628, 122 S.Ct. 2450, 153 L.Ed.2d 586 (2002); see also Bender v. Williamsport Area Sch. Dist., 475 U.S. 534, 542, 106 S.Ct. 1326, 89 L.Ed.2d 501 (1986); Brown v. Keene, 33 U.S. (8 Pet.) 112, 116, 8 L.Ed. 885 (1834). We therefore, for purposes of this opinion, have jurisdiction under 28 U.S.C. § 1291. Our review of the jurisdictional questions at issue is plenary. In re NFL Players Concussion Injury Litig., 775 F.3d 570, 576 (3d Cir. 2014).

IV

Like all other federal courts, we are a court of limited jurisdiction, possessing “only that power authorized by Constitution and statute.” Kokkonen v. Guardian Life Ins. Co. of Am., 511 U.S. 375, 377, 114 S.Ct. 1673, 128 L.Ed.2d 391 (1994). As an Article III court established by Congress, our appellate jurisdiction is “purely statutory.” Heike v. United States, 217 U.S. 423, 428, 30 S.Ct. 539, 54 L.Ed. 821 (1910).

The United States Courts of Appeals have general appellate jurisdiction over “appeals from all final decisions of the district courts of the United States.” 28 U.S.C. § 1291. But carved out of § 1291’s jurisdictional grant is the Court of Appeals for the Federal Circuit. Congress vested that court with “exclusive jurisdiction of an appeal from a final decision of a district court of the United States ... in any civil action arising under ... any Act of Congress relating to patents.” Id. § 1295(a)(1) (emphasis added). The federal district courts, in turn, “have original jurisdiction of any civil action arising under any Act of Congress relating to patents.” Id. § 1338(a). “Thus, the Federal Circuit’s jurisdiction is fixed with reference to that of the district court, and turns on whether the action arises under federal patent law.” Holmes Grp., Inc. v. Vornado Air Circulation Sys., Inc., 535 U.S. 826, 829, 122 S.Ct. 1889, 153 L.Ed.2d 13 (2002). So if the District Court here had jurisdiction over at least one claim in a particular case under § 1338(a), the Federal Circuit has exclusive jurisdiction of that appeal. See Apotex, Inc. v. Thompson, 347 F.3d 1335, 1342 (Fed. Cir. 2003); see also 19 James Wm. Moore & George C. Pratt, Moore’s Federal Practice § 208.10[2], p. 208-16 (3d ed. 2017) (“The minimum jurisdictional requirement is the existence of at least one claim under the patent ... statutes, and in a mixed case, the Federal Circuit has jurisdiction to decide all of the issues involved in the appeal.” (footnote omitted)). In that circumstance, we would lack jurisdiction and be required to transfer these appeals to the Federal Circuit. See 28 U.S.C. § 1631; In re Arunachalam, 812 F.3d 290, 293-94 (3d Cir. 2016) (per curiam).

The discussion that follows applies to both sets of appeals. Consequently, unless otherwise indicated, we will refer to the Lipitor and Effexor plaintiffs collectively as the “plaintiffs” and the Lipitor and Effexor defendants collectively as the “defendants.”

The Supreme Court’s pathmarking decision addressing the Federal Circuit’s patent-law jurisdiction is Christianson v. Colt Industries Operating Corp., 486 U.S. 800, 108 S.Ct. 2166, 100 L.Ed.2d 811 (1988). At the time, the Federal Circuit’s jurisdictional statute vested that court with “exclusive jurisdiction of an appeal from a final decision óf a district court of the United States ... if the decision of a district court was based, in whole or in part, on [28 U.S.C.] § 1338.” 28 U.S.C. § 1295(a)(1). Then, as now, § 1338(a) granted the district courts “original jurisdiction of any civil action arising under any Act of Congress relating to patents.” Section 1338(a) uses the same operative language as 28 U.S.C. § 1331, the statute that gives the district courts “original jurisdiction of all civil actions arising under the Constitution, laws, or treaties of the United States.” (Emphasis added.).

Christianson held that “[l]inguistic consistency” requires that courts apply the same jurisdictional test to determine whether a case arises under § 1331 as it would under § 1338(a). 486 U.S. at 808, 108 S.Ct. 2166. Under § 1338(a), then, jurisdiction extends “only to those cases in which a well-pleaded complaint establishes either that federal patent law creates the cause of action or that the plaintiffs right to relief necessarily depends on resolution of a substantial question of federal patent law, in that patent law is a necessary element of one of the well-pleaded claims.” Id. at 809, 108 S.Ct. 2166. As in the § 1331 context, the determination whether a claim “arises under” patent law must be made in accordance with the time-honored well-pleaded-complaint rule. And as “appropriately adapted to § 1338(a),” that rule provides that the answer to whether a claim “arises under” patent law “must be determined from what necessarily appears in the plaintiffs statement of his own claim in the bill or declaration, unaided by anything alleged in anticipation or avoidance of defenses which it is thought the defendant may interpose.” Id. (quoting Franchise Tax Bd. of Cal. v. Constr. Laborers Vaca tion Trust, 463 U.S. 1, 10, 103 S.Ct. 2841, 77 L.Ed.2d 420 (1983)).

For those cases in which federal patent law does not create the cause of action, it is not “necessarily sufficient that a well-pleaded claim alleges a single theory under which resolution of a patent-law question is essential.” Id. at 810, 108 S.Ct. 2166. Rather, if “ ‘on the face of a well-pleaded complaint there are ... reasons completely unrelated to the provisions and purposes of [the patent laws] why the [plaintiff] may or may not be entitled to the relief it seeks,’ then the claim does not ‘arise under’ those laws.” Id. (alterations in original) (quoting Franchise Tax Bd., 463 U.S. at 26, 103 S.Ct. 2841). “Thus,” Christianson explained, “a claim supported by alternative theories in the complaint may not form the basis for § 1338(a) jurisdiction unless patent law is essential to each of those theories.” Id.

The complaint in Christianson contained an antitrust count that the Court understood as raising a monopolization claim under section 2 of the Sherman Act and a group-boycott claim under section 1. See id. Even though the claims included allegations of patent invalidity, the Court held that the Federal Circuit lacked jurisdiction because the “patent-law issue, while arguably necessary to at least one theory under each claim, [was] not necessary to the overall success of either claim.” Id.

As to the complaint’s section 2 monopolization claim, the Court first identified the “thrust” of the allegations, namely, that Colt, the defendant, “embarked on a course of conduct to illegally extend its monopoly position with respect to the described patents and to prevent” plaintiffs from competing. Id. But because the well-pleaded-eomplaint rule “focuses on claims, not theories,” the Court emphasized that “just because an element that is essential to a particular theory might be governed by federal patent law does not mean that the entire monopolization claim ‘arises under’ patent law.” Id. at 811, 108 S.Ct. 2166. One such theory involved allegations that certain Colt trade secrets were not protected under state law because their underlying patents were invalid. But after parsing the complaint, the Court observed that this monopolization theory was “only one of several, and the only one for which the patent-law issue is even arguably essential.” Id. Because there were “ ‘reasons completely unrelated to the provisions and purposes’ of federal patent law why [the plaintiffs] ‘may or may not be entitled to the relief they [sought]’ under their monopolization claim, the claim [did] not ‘arise under’ patent law.” Id. at 812, 108 S.Ct. 2166 (quoting Franchise Tax Bd., 463 U.S. at 26, 103 S.Ct. 2841).

The same result obtained with regard to the plaintiffs’ section 1 group-boycott claim. That claim involved allegations that Colt engaged in a group-boycott to protect its trade secrets. And like the section 2 monopolization claim, one theory of recovery involved assertions that Colt’s patents protecting its trade secrets were invalid. “Whether or not the patent-law issue was an ‘essential’ element of that group-boycott theory,” the Court noted, plaintiffs “could have supported their group-boycott claim with any of several theories having nothing to do with the validity of Colt’s patents.” Id. at 813, 108 S.Ct. 2166. Instead, “the appearance on the complaint’s face of an alternative, non-patent theory compelled] the conclusion that the group-boycott claim [did] not ‘arise under’ patent law.” Id.

Four working principles underlie the Court’s decision in Christianson. First, whether a claim “arises under” federal patent law is made by reference to the well-pleaded complaint. See Holmes Grp., 535 U.S. at 829-30, 122 S.Ct. 1889. Second, for jurisdictional purposes, regardless of how a complaint labels its claims or counts, courts are to look to the complaint and its allegations as a whole to identify the plaintiffs claims and any theories undergirding those claims. Third, in the antitrust context, courts must attend to the thrust of the plaintiffs allegations and then determine the theories that explain why certain alleged conduct was anticompetitive. And finally, after distinguishing between claims and theories, courts then must ascertain whether each theory supporting a claim necessarily requires the resolution of a substantial question of patent law. If one theory does not, the Federal Circuit lacks appellate jurisdiction. See ClearPlay, Inc. v. Abecassis, 602 F.3d 1364, 1369 (Fed. Cir. 2010) (“Christianson embraces a distinctly non-holistic approach to ‘arising under’ jurisdiction. It is not enough that patent law issues are in the air. Instead, resolution of a patent law issue must be necessary to every theory of relief under at least one claim in the plaintiffs complaint.” (emphasis added)).

Applying these principles, we conclude that the actions brought by the Lipitor and Effexor plaintiffs do not “arise under” patent law. We note at the outset a clear and undisputed aspect of our jurisdictional inquiry. Federal and state antitrust law, not federal patent law, creates plaintiffs’ claims. This case, like Christianson itself, turns on the second head of “arising under” jurisdiction. And so we must decide whether plaintiffs’ well-pleaded complaints state at least one claim upon which their “right to relief necessarily depends on resolution of a substantial question of federal patent law, in that patent law is a necessary element of one of the well-pleaded claims.” Christianson, 486 U.S. at 809, 108 S.Ct. 2166.

Defendants do not argue that plaintiffs’ section 1 restraint of trade claims arise under patent law. Those claims relate only to the Pfizer-Ranbaxy and Wyeth-Teva reverse-payment settlements. Defendants instead home in on plaintiffs’ section 2 monopolization claims. Recall that the thrust of those claims is that Pfizer and Wyeth each engaged in an overall scheme to monopolize the markets for their respective branded Lipitor and Effexor XR drugs. Those schemes, plaintiffs allege, were furthered in part by the companies’ fraudulent procurement and enforcement of certain patents relating to the drugs. But the schemes were also furthered by the-reverse-payment settlements (and, in the Lipitor appeals, the filing of a sham FDA citizen petition).

The fraudulent procurement of a patent—known as Walker Process fraud, see Walker Process Equip., Inc. v. Food Mach. & Chem. Corp., 382 U.S. 172, 86 S.Ct. 347, 15 L.Ed.2d 247 (1965) (recognizing that a patentee’s knowing and willful misrepresentation of facts to the PTO can strip the patentee of immunity under the antitrust laws) — requires a plaintiff to show, among other things, that the paten-tee committed fraud before the PTO, that the fraud caused the patent to issue, and that the patentee enforced the fraudulently procured patent, Unitherm Food Sys., Inc. v. Swift-Eckrich, Inc., 375 F.3d 1341, 1355 (Fed. Cir. 2004), rev’d on other grounds, 546 U.S. 394, 126 S.Ct. 980, 163 L.Ed.2d 974 (2006). Walker Process fraud has for some time been considered by courts to present a substantial question of patent law. See In re DDAVP Antitrust Litig., 585 F.3d 677, 685 (2d Cir. 2009); In re Ciprofloxacin Antitrust Litig., 544 F.3d 1323, 1330 n.8 (Fed. Cir. 2008) (“[T]he determination of fraud before the PTO necessarily involves a substantial question of patent law.”); Nobelpharma AB v. Im plant Innovations, Inc., 141 F.3d 1059, 1068 (Fed. Cir. 1998) (en banc in relevant part) (“[WJhether conduct in procuring or enforcing a patent is sufficient to strip a patentee of its immunity from the antitrust laws is to be decided as a question of Federal Circuit law.”). And to the extent plaintiffs’ sham litigation and false Orange Book listing theories depend on a successful showing of Walker Process fraud, they too could present substantial questions of patent law. See DDAVP, 585 F.3d at 685; Nobelpharma, 141 F.3d at 1071-72. We recognize as well that the substantiality of these theories may be open to debate following Gunn v. Minton, 568 U.S. 251, 133 S.Ct. 1059, 185 L.Ed.2d 72 (2013). That case held, in the context of a state legal malpractice claim, that hypothetical, backward-looking, case-within-a-case questions of patent law that do not change the real-world result of prior federal patent litigation do not present a substantial patent-law issue. Id. > at 1067-68. We need not definitively address the substantiality of plaintiffs’ Walker Process, sham litigation, and false Orange Book listing theories in light of Gunn. For even assuming that these theories do present substantial questions of patent law, plaintiffs’ right to relief on their section 2 monopolization claims does not depend upon them.

Here, plaintiffs could obtain relief on their section 2 monopolization claims by prevailing on an alternative, non-patent-law theory, namely, that Pfizer and Wyeth monopolized the market in their respective branded drugs by engaging in a reverse-payment settlement. And in Lipitor the plaintiffs could also prevail on the additional non-patent law theory that Pfizer filed a sham citizen petition with the FDA. See DDAVP, 585 F.3d at 686 (“[WJhether [a FDA] petition was a sham is an issue independent of patent law.”); see also Apo-tex Inc. v. Acorda Therapeutics, Inc., 823 F.3d 51, 59 (2d Cir. 2016).

Actavis teaches that reverse-payment antitrust claims do not present a question of patent law. See 133 S.Ct. at 2236-37 (“[T]he size of the unexplained reverse payment can provide a workable surrogate for a patent’s weakness, all without forcing a court to conduct a detailed exploration of the validity of the patent itself.”). The Court did acknowledge, however, that questions of patent validity may still arise from time to time. See id. at 2236 (“[IJt is normally not necessary to litigate patent validity to answer the antitrust question (unless, perhaps, to determine whether the patent litigation is a sham).”). But even where patent-law questions are presented, it does not follow that patent law is necessary for relief on every theory of liability supporting an antitrust claim. In the present appeals, “[sjince there are reasons completely unrelated to the provisions and purposes of federal patent law why [plaintiffs] may or may not be entitled to the relief they seek under their monopolization claim, the claim does not ‘arise under’ federal patent law.” Christianson, 486 U.S. at 812, 108 S.Ct. 2166 (brackets, citation, and some internal quotation marks omitted). These considerations lead us to conclude that the presence of non-patent-law theories of liability supporting the Lipitor and Effexor plaintiffs’ monopolization claims vests jurisdiction over their appeals in this Court, not the Federal Circuit.

C

Defendants do not quarrel with any of the principles that guide our analysis. They instead assert that plaintiffs’ reverse-payment settlement allegations constitute monopolization claims separate and apart from the Walker ■Process fraud, sham litigation, and false Orange Book listing theories. The allegations of fraudu lent procurement and enforcement of the Lipitor and Effexor patents, in defendants’ view, involve distinct anticompetitive conduct that occurred years before the reverse-payment settlements (and, in Lipitor, the sham FDA citizen petition).

We reject this divide-and-conquer approach to “arising under” jurisdiction. Defendants in effect ask that we rewrite plaintiffs’ complaints, which plead patent-law related theories as aspects of an overall monopolistic scheme. A monopolization claim under section 2 of the Sherman Act has two elements: (1) the possession of monopoly power in the relevant market and (2) the willful acquisition or maintenance of that power. LePage’s Inc. v. 3M, 324 F.3d 141, 146 (3d Cir. 2003) (en banc) (citing United States v. Grinnell Corp., 384 U.S. 563, 570-71, 86 S.Ct. 1698, 16 L.Ed.2d 778 (1966)). But to be con demned as exclusionary, a monopolist’s anticompetitive conduct must have an anti-competitive effect. “The relevant inquiry,” we have held, “is the anticompetitive effect of [a defendant’s] exclusionary practices considered together.” Id. at 162. Thus, “courts must look to the monopolist’s conduct taken as a whole rather than considering each aspect in isolation.” Id. (citing Cont’l Ore Co. v. Union Carbide & Carbon Corp., 370 U.S. 690, 699, 82 S.Ct. 1404, 8 L.Ed.2d 777 (1962)); see id. (“[I]t would not be proper to focus on specific individual acts of an accused monopolist while refusing to consider their overall combined effect.... We are. dealing with what 'has been called the ‘synergistic effect’ of the mixture of the elements.” (alterations in original) (internal quotation marks omitted)).

Defendants contend that the patent-law theories of monopolization liability in plaintiffs’ complaints are distinct “claims.” But that runs headlong into traditional antitrust principles. Plaintiffs’ monopolization claims encompass the totality of the allegedly anticompetitive conduct — from defendants’ fraudulent procurement and enforcement of their patents on through to the reverse-payment settlements. We will not permit the defendants to commandeer these complaints, of which plaintiffs are master.

Nor do we accept the argument that certain statements made by the Effe-xor plaintiffs in the District Court somehow estop them from arguing that the patent-law allegations constitute theories of relief. Principles of estoppel cannot confer jurisdiction where it otherwise does not exist. See Ins. Corp. of Ireland, Ltd. v. Compagnie des Bauxites de Guinee, 456 U.S. 694, 702, 102 S.Ct. 2099, 72 L.Ed.2d 492 (1982); Semper v. Gomez, 747 F.3d 229, 247 (3d Cir. 2014). And in any event, our jurisdictional inquiry is confined solely to the plaintiffs’ well-pleaded complaints, not subsequent events. See Christianson, 486 U.S. at 814, 108 S.Ct. 2166 (“Since the district court’s jurisdiction is determined by reference to the well-pleaded complaint, not the well-tried case, the referent for the Federal Circuit’s jurisdiction must be the same.”).

D

Our jurisdictional holding is consistent, we think, with two of the Second Circuit’s pre-Actavis cases. In one case, the court transferred an appeal to the Federal Circuit and retained jurisdiction over others. The Second Circuit explained: “The indirect purchaser plaintiffs amended their complaint to add state-law, Walker Process antitrust claims.... Because the Walker Process claims are preempted by patent law, we transferred the indirect purchaser plaintiffs’ appeal to the Federal Circuit, while retaining jurisdiction over the direct purchaser plaintiffs’ appeals.” Arkansas Carpenters Health & Welfare v. Bayer AG, 604 F.3d 98, 103 n.10 (2d Cir. 2010) (per curiam); see In re Ciprofloxacin Hydrochloride Antitrust Litig., 2007 U.S. App. LEXIS 30732, at *1 (2d Cir. Nov. 7, 2007) (order transferring indirect purchaser plaintiffs’ appeal to the Federal Circuit). The Second Circuit and the Federal Circuit therefore each independently assessed the lawfulness of the same reverse-payment settlement. See Arkansas Carpenters, 604 F.3d at 103 & n.10; Ciprofloxacin, 544 F.3d at 1333. But unlike the Lipitor and Effexor appeals before us, the appeal transferred from the Second Circuit to the Federal Circuit involved stand-alone Walker Process claims. See In re Ciprofloxacin Hydrochloride Antitrust Litig., 363 F.Supp.2d 514, 544 (E.D.N.Y. 2005) (“Kin-direct plaintiffs’ Count V [raising state-law Walker Process claims] not only arises out of patent law, but rests entirely on patent law” (emphasis added)), aff'd, 544 F.3d 1323 (Fed. Cir. 2008), and aff'd sub nom. Arkansas Carpenters, 604 F.3d 98.

And in DDAVP, 585 F.3d 677, the Second Circuit retained jurisdiction over an antitrust case presenting patent-law issues. The DDAVP plaintiffs alleged four theories of liability in a Sherman Act monopolization claim against a branded drug manufacturer based upon theories nearly identical to those the Lipitor and Effexor plaintiffs bring against Pfizer and Wyeth: Walker Process fraud, sham Orange Book listing, sham litigation against generic competitors, and a sham FDA citizen petition. Id. at 685. The Second Circuit acknowledged that, while the plaintiffs’ first three theories turned on substantial questions of patent law, the fourth theory — the filing of a sham FDA citizen petition — did not. Id. at 685-86. Because the citizen-petition theory did not raise any question of patent law, the court exercised jurisdiction over the entirety of the plaintiffs’ monopolization claim. Id. at 686.

A final, prudential consideration tips in favor of our Court exercising jurisdiction over these appeals. Under the Federal Circuit’s choice-of-law rules, it would apply Third Circuit antitrust jurisprudence — including our recent decision in King Drag, 791 F.3d 388—when reviewing whether plaintiffs’ complaints state plausible claims for relief under Actavis. See Nobelpharma, 141 F.3d at 1059 (Federal Circuit “appl[ies] the law of the appropriate regional circuit to issues involving other elements of antitrust law such as relevant market, market power, damages, - etc., as those issues are not unique to patent law”). Now that the Supreme Court has confirmed that it is usually unnecessary to litigate these patent-law issues to determine antitrust liability, the development of post-AcidTOS jurisprudence is, in the ordinary case, left to the regional Courts of Appeals.

Christianson establishes that not all cases involving patent law fall within the Federal Circuit’s jurisdiction. Congress has left a role for our Court to play in adjudicating patent-law issues over which we possess jurisdiction. Our holding requires us to fulfill that role in these appeals.

V

The appeal of the RP Healthcare plaintiffs requires a separate jurisdictional inquiry. That case was filed by a group of California pharmacists in the Superior Court of California, Sonoma County, but Pfizer removed it to federal district court, citing federal-question jurisdiction under 28 U.S.C. § 1331 and patent-law jurisdiction under § 1338(a). RP Healthcare J.A. 26-27; see 28 U.S.C. § 1441(a). In denying the RP Healthcare plaintiffs’ remand motion, the District Court reasoned that “there may be patent issues raised as defenses in this case which would engender jurisdiction.” Lipitor J.A. 2. We disagree. “Under the well-pleaded complaint rule ... whether a claim ‘arises under’ patent law ‘must be determined from what necessarily appears in the plaintiffs statement of his own claim in the bill or declaration, unaided by anything alleged in anticipation or avoidance of defenses which it is thought the defendant may interpose.’” Christianson, 486 U.S. at 809, 108 S.Ct. 2166 (quoting Franchise Tax Bd., 463 U.S. at 8, 103 S.Ct. 2841); see Louisville & Nashville R.R. Co. v. Mottley, 211 U.S. 149, 29 S.Ct. 42, 53 L.Ed. 126 (1908); N.J. Carpenters v. Tishman Constr. Corp. of N.J., 760 F.3d 297, 302 (3d Cir. 2014) (“The existence or expectation of a federal defense is insufficient to confer federal jurisdiction.”).

Pfizer and Ranbaxy nevertheless argue that the RP Healthcare case belongs in federal court because it “arises under” patent law pursuant to. § 1338(a). They also say the District Court possessed diversity jurisdiction before final judgment entered as a result of the RP Healthcare plaintiffs’ voluntary dismissal of the only two non-diverse defendants. We reject the first argument but find the record insufficient to decide the second.

The RP Healthcare plaintiffs do not challenge the Pfizer-Ranbaxy settlement as an unlawful reverse payment. Rather, they allege that the settlement constitutes a per se unlawful market allocation agreement in violation of California’s Cartwright Act. Two years after Actavis, the California Supreme Court held that reverse-payment settlements can be challenged under that Act and are to be analyzed under a structured rule-of-reason. In re Cipro Cases I & II, 61 Cal.4th 116, 187 Cal. Rptr.3d 632, 348 P.3d 845 (2015). But the California court has yet to recognize the kind of per se market allocation claim proposed by the RP Healthcare plaintiffs.

To the extent their claim exists under California law (a question we do not decide), as pled by the RP Healthcare plaintiffs that claim would not “arise under” federal patent law. Pfizer and Ran-baxy latch onto a single sentence in the RP Healthcare plaintiffs’ state court complaint making an express allegation of Walker Process fraud. See RP Healthcare Pls.’ Compl. ¶ 114, RP Healthcare J.A. 57 (“The Agreement between Defendants extending the length of the Lipitor patents constitutes fraudulent procurement and enforcement of a patent....” (citing Walker Process, 382 U.S. 172, 86 S.Ct. 347, 15 L.Ed.2d 247)). But like the complaints of the Lipitor and Effexor plaintiffs discussed above, we conclude that there are alternative non-patent-law theories through which the RP Healthcare plaintiffs could prevail on their state-law antitrust claim. See Christianson, 486 U.S. at 809-10, 108 S.Ct. 2166. The RP Healthcare plaintiffs’ complaint includes theories of liability other than Walker Process fraud. See id. ¶ 105, RP Healthcare J.A. 56 (“The Agreements between the Defendants, which artificially extended the length of the Lipitor-related patents, allocated markets between them, artificially postponed price reductions, and restrained trade in the provision of Lipitor and its generic alternatives, are a violation of the Cartwright Act....” (emphasis added)). Thus, the RP Healthcare plaintiffs could obtain relief on the market allocation claim all without addressing the validity of Pfizer’s Lipitor patents. The oblique mention of Walker Process fraud in their complaint does not land this case in the “special and small category” of state-law claims “in which arising under jurisdiction still lies.” Gunn, 133 S.Ct. at 1064 (internal quotation marks omitted).

B

While the District Court did not possess jurisdiction over the RP Healthcare case under § 1338(a), the possibility exists that the court had diversity jurisdiction by the time it entered final judgment. Article III of the Constitution provides that “[t]he judicial Power [of the United States] shall extend ... to Controversies ... between Citizens of different States; ... and between a State, or the Citizens thereof, and foreign States, Citizens or Subjects.” Beginning with the Judiciary Act of 1789, ch. 20, § 11, 1 Stat. 78, Congress has authorized the federal courts to exercise jurisdiction based on the parties’ diversity of citizenship. In its current form, the diversity statute vests in the federal district courts original jurisdiction of “all civil actions where the matter in controversy exceeds the sum or value of $75,000, ... and is between ... citizens of different States and in which citizens or subjects of a foreign state are additional parties.” 28 U.S.C. § 1132(a)(3). Since Strawbridge v. Curtiss, 7 U.S. (3 Cranch) 267, 2 L.Ed. 435 (1806), the Supreme Court has interpreted the diversity statute to require “complete diversity” of citizenship: “[i]n a case with multiple plaintiffs and multiple defendants, the presence in the action of a single plaintiff from the same State as a single defendant deprives the district court of original diversity jurisdiction over the entire action,” Exxon Mobil Corp. v. Allapattah Servs., Inc., 545 U.S. 546, 553, 125 S.Ct. 2611, 162 L.Ed.2d 502 (2005).

Though “[i]t had long been the case that ‘the jurisdiction of the court depends upon the state of things at the time of the action brought,’ ” Grupo Dataflux v. Atlas Global Grp., L.P., 541 U.S. 567, 570, 124 S.Ct. 1920, 158 L.Ed.2d 866 (2004) (quoting Mollan v. Torrance, 22 U.S. (9 Wheat.) 537, 539, 6 L.Ed. 154 (1824)), this time-of-filing rule is subject to a few discrete exceptions. One such “method of curing a jurisdictional defect [that has] long been an exception to the time-of-filing rule” is when a jurisdictional defect is “cured by the dismissal of the party that had destroyed diversity.” Id. at 572, 124 S.Ct. 1920. As the Supreme Court recognized in Caterpillar Inc. v. Lewis, “a district court’s error in failing to remand a case improperly removed is not fatal to the ensuing adjudication if federal jurisdictional requirements are met at the time judgment is entered.” 519 U.S. 61, 64, 117 S.Ct. 467, 136 L.Ed.2d 437 (1996).

Pfizer and Ranbaxy urge us to apply that exception here. After all, the RP Healthcare plaintiffs voluntarily dismissed the only two non-diverse defendants prior to entry of final judgment. Before this Court, however, the parties expressed uncertainty regarding the state of the record as it pertains to the citizenship of two parties- — defendants Pfizer Ireland Pharmaceuticals and Warner-Lambert Co., LLC, both unincorporated entities and wholly owned subsidiaries of Pfizer. See Lipitor Tr. of Oral Arg. 23-24, 44-47; RP Healthcare Pls.’ Reply Br. 17-18. Like all unincorporated entities, partnerships and limited liability companies (LLCs) bear the citizenship of each of their members. See Americold Realty Trust v. ConAgra Foods, Inc., — U.S. -, 136 S.Ct. 1012, 1016-17, 194 L.Ed.2d 71 (2016); Carden v. Arkoma Assocs., 494 U.S. 185, 195-96, 110 S.Ct. 1015, 108 L.Ed.2d 157 (1990); Zambelli Fireworks Mfg. Co. v. Wood, 592 F.3d 412, 420 (3d Cir. 2010).

As the parties asserting diversity jurisdiction, Pfizer and Ranbaxy bear the burden of proving diversity of citizenship by a preponderance of the evidence. See Freidrich v. Davis, 767 F.3d 374, 377 (3d Cir. 2014). Since this case was removed to federal court, diversity must have exist ed both at the time the RP Healthcare plaintiffs’ state court complaint was filed and at the time of removal. See Pullman Co. v. Jenkins, 305 U.S. 534, 537, 59 S.Ct. 347, 83 L.Ed. 334 (1939); Johnson v. SmithKline Beecham Corp., 724 F.3d 337, 346 (3d Cir. 2013). But no changes in citizenship after the time of filing (and, as relevant here, the time of removal) can create or destroy diversity. See Grupo Dataflux, 541 U.S. at 574-75, 124 S.Ct. 1920; Conolly v. Taylor, 27 U.S. (2 Pet.) 556, 565, 7 L.Ed. 518 (1829).

In calling for diversity jurisdiction Pfizer and Ranbaxy made no effort before this Court or the District Court to demonstrate that complete diversity was in fact present before final judgment. That is especially puzzling, since an unincorporated association “is in the best position to ascertain its own membership,” Lincoln Benefit Life Co. v. AEI Life, LLC, 800 F.3d 99, 108 (3d Cir. 2015), and the entities in question are Pfizer subsidiaries. While we have previously observed that, “where the unincorporated association is the proponent of diversity jurisdiction, there is no reason to excuse it of its obligation to plead the citizenship of each of its members,” id. at 108 n.36, that statement was made in the context of an unincorporated association asserting diversity as a plaintiff. It does not address the situation in this ease, where the removing parties are asserting diversity as a result of the plaintiffs’ own voluntary post-removal actions. We therefore consider it premature to direct that the RP Healthcare case be sent back to California state court. Rather, we will remand the matter to the District Court to give the parties the opportunity to clarify the record with regard to diversity of citizenship. The District Court should also ensure that the amount in controversy alleged in the RP Healthcare plaintiffs’ state-court complaint exceeds $75,000. See 28 U.S.C. § 1332(a); Angus v. Shiley, 989 F.2d 142, 145-46 (3d Cir. 1993).

Our remand applies as well to the Daiiehi Sankyo defendants. Before the District Court, they moved to dismiss the RP Healthcare plaintiffs’ complaint on three grounds: lack of Article III standing, lack of personal jurisdiction, and failure to state a claim upon which relief can be granted. The District Court dismissed the Daiiehi Sankyo defendants under Rule 12(b)(6) for failure to state a plausible claim. Lipitor J.A. 65, 3543-44. But “a federal court generally may not rule on the merits of a case without first determining that it has jurisdiction over the category of claim in suit (subject-matter jurisdiction) and the parties (personal jurisdiction).” Sinochem Int’l Co. v. Malaysia Int’l Shipping Corp., 549 U.S. 422, 430-31, 127 S.Ct. 1184, 167 L.Ed.2d 15 (2007); see Steel Co. v. Citizens for Better Environment, 523 U.S. 83, 93-102, 118 S.Ct. 1003, 140 L.Ed.2d 210 (1998). The District Court should have resolved the standing and personal jurisdictional arguments before dismissing Daiiehi Sankyo on the merits. In the event that the District Court concludes on remand that the parties were completely diverse at the time of judgment, it should address those arguments to determine whether it had the power to reach the merits of the RP Healthcare plaintiffs’ claim against Daiiehi Sankyo.

It is a common practice among the Courts of Appeals to retain jurisdiction over an appeal while making a limited remand for additional findings or explanations. Basic illustrations include a “controlled remand to determine whether there is federal subject-matter jurisdiction,” as well as “remands to determine justiciability or personal jurisdiction.” 16 Charles Alan Wright, Arthur R. Miller, & Edward H. Cooper, Federal Practice & Procedure § 3937.1, pp. 847-48 (3d ed. 2012) (footnote omitted); see, e.g., Friery v. Los Angeles Unified Sch. Dist., 448 F.3d 1146, 1150 (9th Cir. 2006) (limited remand for Article III standing determination); Fort Knox Music Inc. v. Baptiste, 203 F.3d 193, 197 (2d Cir. 2000) (limited remand for personal jurisdiction determination); Jason’s Foods, Inc. v. Peter Eckrich & Sons, Inc., 768 F.2d 189, 190-91 (7th Cir. 1985) (limited remand for diversity-of-citizenship determination). We will follow that practice and retain jurisdiction over the RP Healthcare plaintiffs’ appeal. It is expected that the' District Court and the parties will move expeditiously on remand to resolve the diversity-of-citizenship issue and, if necessary, jurisdiction over the Daiichi Sankyo defendants.

VI

For the reasons stated, we conclude that, with a single exception, we have jurisdiction to reach the merits of these appeals. In one of the Lipitor appeals, RP Healthcare, Inc. v. Pfizer, Inc., No. 14-4632, because it is unclear whether the District Court had jurisdiction at the time judgment was entered, we will order a limited remand for the parties to clarify the record in this regard. Any further proceedings in these appeals will be heard by this panel.