In Re Glumetza Antitrust Litigation

District Court, N.D. California·Decided August 15, 2020·No. 3:19-cv-05822·Unknown

Opinion

1 2 3 4 5 UNITED STATES DISTRICT COURT 6 NORTHERN DISTRICT OF CALIFORNIA 7

8 In re No. C 19-05822 WHA 9 GLUMETZA ANTITRUST No. C 19-06138 WHA LITIGATION. No. C 19-06839 WHA 10 No. C 19-07843 WHA No. C 19-08155 WHA 11 No. C 20-01198 WHA This Document Relates to: No. C 20-05251 WHA 12 DIRECT PURCHASER ACTIONS. (Consolidated) 13 ORDER CERTIFYING CLASS 14 15 INTRODUCTION 16 In this antitrust action arising from an alleged reverse-payment settlement of a patent 17 infringement suit between brand and generic marketers of the diabetes drug Glumetza, direct 18 purchaser plaintiffs move for class certification. Common issues predominate, the putative 19 class’s market-impact and damages models adequately reflect the theory of liability, and any 20 assumptions underlying the model either go to the heart of the merits or can be modified to 21 account for a range of jury determinations. A class is CERTIFIED. 22 STATEMENT 23 A prior order details this case (Dkt. No. 188 as amended Dkt. No. 204). In re Glumetza 24 Antitrust Litigation, ___ F. Supp. 3d ___, 2020 WL 1066934 (N.D. Cal. Mar. 5, 2020). But the 25 essence bears restating. This case arises from a perversion of the patent and pharmaceutical- 26 regulatory framework. See generally 35 U.S.C. § 156 et seq., 21 U.S.C. § 301 et seq. The 27 Hatch-Waxman Act implements a network of incentives to encourage faster introduction of 1 ease FDA approval, generic manufacturers may file an Abbreviated New Drug Application to 2 piggyback on the approval process for the underlying brand drug. See FTC v. Actavis, 570 U.S. 3 136, 142 (2013); 21 U.S.C. §§ 355(j)(2)(A)(ii), (iv). 4 Patents covering the brand drug can still spoil the fun, though. To gain approval, an 5 ANDA applicant must certify to the FDA that no brand patents block the generic drug’s market 6 entry. If the brand holds live patents, for example, the generic must file a “Paragraph IV 7 certification” of noninfringement or invalidity. Even so, if the brand manufacturer promptly 8 sues for infringement, the FDA can’t approve the generic for 30 months (or until the end of the 9 suit, whichever comes first). Actavis, 570 U.S. at 143; 21 U.S.C. §§ 355(j)(2)(A)(vii)(I)–(IV), 10 (5)(B)(iii); 35 U.S.C. § 271(e)(2)(A). 11 The Hatch-Waxman scheme encourages patent-challenge certifications by granting 180 12 days of generic market exclusivity to the first generic to file such an application. If this “first 13 filer” wins the infringement suit and markets, it gets 180 days to compete alone with the brand 14 drug, meaning the FDA can’t approve any other generics during that time. This can be “worth 15 several hundred million dollars” to the generic manufacturer and outweigh the risk of 16 infringement suit. Actavis, 570 U.S. at 143–44; 21 U.S.C. § 355(j)(5)(B)(iv). But this 180-day 17 exclusivity period doesn’t stop the brand manufacturer from marketing an “authorized generic” 18 to recoup some of those millions. See Teva Pharm. v. Crawford, 410 F.3d 51, 55 (D.C. Cir. 19 2005). Moreover, the first filer can forfeit the 180-day exclusivity if it stalls too long. 21 20 U.S.C. § 355(j)(5)(D)(i)(I)(aa), (iii). 21 This scheme is supposed to get us faster, cheaper generic drugs. But the industry found a 22 way to do the opposite. Sometimes a brand drug manufacturer sues an ANDA filer, and the 23 brand manufacturer pays the generic to settle. In exchange, the first filer generic manufacturer 24 agrees to stay off the market for a few years. Instead of expedited generic entry, the brand 25 maintains its monopoly and cuts the supposed-generic a share of the profits. In 2013, the 26 United States Supreme Court found that these “pay for delay” schemes can violate federal 27 antitrust law. Actavis, 570 U.S. at 158–60. 1 Our putative class alleges such a scheme involving the diabetes medication Glumetza. 2 But instead of a cash payment, our brand manufacturer gave something else of value, a virtual 3 guarantee that our generic manufacturer would face no generic competition (authorized or 4 otherwise) for at least a year after its belated market entry. This allowed the brand 5 manufacturer to then raise the price of pills from $5.72 to over $51 each. 6 Some iteration of defendant brand manufacturer Bausch Health Companies Inc. has 7 marketed Glumetza since 2005. In July 2009, defendant generic manufacturers Lupin 8 Pharmaceuticals, Inc. and Lupin Ltd. filed an ANDA to market generic versions of Glumetza 9 and certified noninfringement or invalidity against the four relevant patents: U.S. Patent Nos. 10 6,340,475; 6,635,280; 6,488,962; and 6,723,340. Assertio Therapeutics, Inc. (then owner of 11 Glumetza marketing) sued Lupin for patent infringement in November 2009, triggering the 30- 12 month stay against FDA approval. When the FDA tentatively approved the ANDA in January 13 2012 (meaning Lupin could market its generic but for the 30-month stay), Assertio and others 14 involved in Glumetza marketing promptly settled with Lupin. See Depomed, Inc. v. Lupin 15 Pharms., Inc., No. C 09-05587 PJH, Dkt. No. 152 (N.D. Cal. Mar. 27, 2012). 16 Under the settlement, Lupin agreed to walk away, leave the patents alone, and not market 17 a generic Glumetza for four years, until February 2016. In return, Assertio and Santarus 18 promised no authorized generic would compete with Lupin for at least a year once its generic 19 entered the market. The settlement also included two clauses to protect Lupin from other 20 generic competition. The “most-favored-entry” clause expressly provided that if any other 21 generic succeeded in marketing a generic Glumetza before February 2016, Lupin could market 22 immediately. Then, the “most-favored-entry-plus” clause stated that Assertio and Santarus 23 would not license any other generic Glumetza manufacturers until at least 180 days (though our 24 facts seem to indicate a full year) following Lupin’s market entry. These provisions undercut 25 the incentive for any other generic manufacturer to enter the market before Lupin. 26 But only the first of these terms, that Lupin would not market until February 2016, made it 27 into the parties’ stipulated dismissal. Curiously, or as the putative class alleges, deliberately, 1 the no-authorized generic, the most-favored-entry, or the most-favored-entry-plus clauses. All 2 of that was in a side agreement. Though defendants contend they put the world on notice of the 3 no-authorized generic provision in July 2015, on these pleadings, defendants did not disclose 4 the provision until February 2016. Glumetza, 2020 WL 1066934 at *7. 5 Apparently the scheme worked. No other generic manufacturers marketed generic 6 Glumetza until well after Lupin. Sun Pharmaceuticals tried, filing its ANDA in May 2011. 7 Asssertio and Santarus promptly sued and a January 2013 settlement allegedly kept Sun’s 8 generic off the market until August 2016. Watson Pharmaceuticals also tried, filing its ANDA 9 in March 2012. Another prompt lawsuit from Assertio and Santarus resulted in a November 10 2013 settlement allegedly keeping Watson’s generic off the market also until August 2016.

Free access — add to your briefcase to read the full text and ask questions with AI

1 2 3 4 5 UNITED STATES DISTRICT COURT 6 NORTHERN DISTRICT OF CALIFORNIA 7

5 All persons or entities in the United States and its territories who directly purchased Glumetza or generic Glumetza from a defendant 6 from May 6, 2012, until the effects of the defendants’ conduct ceased. 7 Defendants oppose. This order follows full briefing and a hearing (held telephonically due to 8 COVID-19). 9 ANALYSIS 10 Numerosity of members, commonality of issues, typicality and adequacy of 11 representatives, and one requirement of Rule 23(b) guard the door to class certification. 12 Abdullah v. U.S. Sec. Assocs., Inc., 731 F.3d 952, 956–57 (9th Cir. 2013). “[Q]uestions of law 13 or fact common to class members [must] predominate over questions affecting only individual 14 members” in direct purchaser plaintiffs’ proposed Rule 23(b)(3) class, which must also be the 15 superior method “for fairly and efficiently adjudicating” this case. Rule 23 doesn’t set forth a 16 mere pleading standard; it demands rigor. A putative class must affirmatively demonstrate that 17 it “in fact” meets these prerequisites. Comcast v. Behrend, 569 U.S. 27, 33 (2013). 18 1. COMMONALITY AND PREDOMINANCE. 19 Commonality requires “questions of law or fact common to the class.” Rule 23(a)(2). “A 20 common contention need not be one that will be answered, on the merits, in favor of the class. 21 It only must be of such nature that it is capable of classwide resolution.” Alcantar v. Hobart 22 Serv., 800 F.3d 1047, 1053 (9th Cir. 2015) (citation and quotation omitted). There need only be 23 “a single significant question of law or fact.” Stockwell v. City & Cty. of San Francisco, 749 24 F.3d 1107, 1111 (9th Cir. 2014). 25 Superseding commonality, predominance under Rule 23(b)(3) asks whether a putative 26 class is “sufficiently cohesive to warrant adjudication by representation.” Predominant 27 questions makeup “a significant aspect of the case” and clearly justify “handling the dispute on 1 a representative rather than on an individual basis.” Predominance isn’t a counting game, 2 though. “Rather, more important questions apt to drive the resolution of the litigation” carry 3 greater weight than less significant individualized questions. So, “even if just one common 4 question predominates, ‘the action may be considered proper under Rule 23(b)(3) even though 5 other important matters will have to be tried separately.’” In re Hyundai & Kia Fuel Econ. 6 Litig., 926 F.3d 539, 557 (9th Cir. 2019) (en banc) (quoting Tyson Foods, Inc. v. Bouaphakeo, 7 577 U.S. ___, 136 S. Ct. 1036, 1045 (2016)). 8 “Merits questions may be considered to the extent — but only to the extent — that they 9 are relevant to determining whether” the putative class has satisfied the requirements for class 10 certification. Amgen Inc. v. Connecticut Retirement Plans & Trust Funds, 568 U.S. 455, 465– 11 66 (2013). That being said, deciding whether common evidence drives class claims necessarily 12 asks what law governs. So our “assessment of predominance begins, of course, with the 13 elements of the underlying cause of action.” Walker v. Life Ins. Co. of the Southwest, 953 F.3d 14 624, 630 (9th Cir. 2020) (quotation omitted). “To establish an antitrust claim, plaintiffs 15 typically must prove (1) a violation of the antitrust laws, (2) an injury they suffered as a result 16 of that violation, and (3) an estimated measure of damages.” In re New Motor Vehicles 17 Canadian Export Antitrust Litig., 522 F.3d 6, 19 n. 18 (1st Cir. 2008)). 18 A. ANTITRUST VIOLATION. 19 Our class invokes both Sections 1 and 2 of the Sherman Act. “Bearing firmly in mind,” 20 however, that Rule 23(b)(3) turns “on the predominance of common questions,” we recall that 21 the question here — one step removed from proof of defendants’ unlawful scheme — remains 22 whether that proof comes common to the class, or whether we will have to search for it class 23 member by class member. See Amgen, 568 U.S. at 466. The Supreme Court has observed that 24 antitrust cases “readily” meet the predominance requirement. See Amchem Prods. v. Windsor, 25 521 U.S. 591, 625 (1997). The underlying purpose of antitrust, to “safeguard general 26 competitive conditions, rather than to protect specific competitors,” illustrates why. The 27 violation turns on defendants’ conduct and intent along with the effect on the market, not on 1 individual class members. See Alaska Airlines v. United Airlines, 948 F.2d 536, 540 (9th Cir. 2 1991). 3 A Section 1 restraint claim requires (1) a contract, combination, or conspiracy that (2) 4 unreasonably restrains (3) interstate commerce. An unreasonable restraint comes either as a per 5 se violation (“when surrounding circumstances make the likelihood of anticompetitive conduct 6 so great as to render unjustified further examination”) or a violation of the rule of reason, 7 meaning the “restraint’s harm to competition outweighs its procompetitive effects.” NCAA v. 8 Bd. of Regents of Univ. Okla., 468 U.S. 85, 103–04 (1984); Tanaka v. Univ. of S. Cal., 252 F.3d 9 1059, 1063 (9th Cir. 2001). In pharmaceutical reverse payment cases, such as this one, the 10 Supreme Court has directed us to apply the rule of reason, explaining that:

11 [T]he likelihood of a reverse payment bringing about anticompetitive effects depends upon its size, its scale in relation to 12 the payor’s anticipated future litigation costs, its independence from other services for which it might represent payment, and the 13 lack of any other convincing justification. 14 Actavis, 570 U.S. at 158–59. Notice then that the illegality of defendants’ scheme turns on 15 details of the payment and defendants’ purposes — evidence common to every purchaser. 16 Section 2 monopolization claims readily lend themselves to common evidence. They 17 require “(1) the possession of monopoly power in the relevant market and (2) the willful 18 acquisition or maintenance of that power as distinguished from growth or development as a 19 consequence of a superior product, business acumen, or historic accident.” Alaska Airlines v. 20 United Airlines, 948 F.2d 536, 541 (9th Cir. 1991) (quoting Aspen Skiing Co. v. Aspen 21 Highlands Skiing Co., 472 U.S. 585, 596 n. 19 (1985)). So, the state of the market and 22 defendants’ use and maintenance of monopoly power, as opposed to individual plaintiff’s 23 conduct, drives the claim. 24 The class confirms its intent to prove the antitrust violation by common proof. In the 25 alleged conspiracy, the generic-drug manufacturer joined, maintained, and partook in the spoils 26 of the brand monopoly. After Assertio and Santarus sued Lupin for seeking FDA approval to 27 market a generic form of Glumetza, they all settled in February 2012. Assertio and Santarus 1 and enjoyed free rein to hike the drug price by nearly 800 percent — from $5.72 to $51 for a 2 single 500mg tablet — in the first half of 2015. Lupin had to sit tight at first. But, in February 3 2016 — free of other generic or authorized generic competition for at least one year — it would 4 hop aboard the cash-cow, pricing its generic slightly below the still-wildly-overpriced brand 5 drug. Though prices would drop as more generics entered the market, due to defendants’ 6 scheme, even today brand and generic Glumetza prices have not yet dropped to pre-hike levels. 7 That is to say that in the class’s telling, defendants wildly succeeded in their antitrust violation, 8 suppressing competition and inflating prices, before we even mention individual class members 9 or, for that matter, the background — the potentially millions of diabetic Americans, their 10 livelihoods dependent upon the prescription drug, faced with the grim choice to either pony up 11 the cash or else. See Alaska Airlines, 948 F.2d at 540. 12 Defendants wisely do not contest that common issues predominate the underlying antitrust 13 violation. Instead, they save their objections for the class’s showing of injury and the measure 14 of damages. 15 B. ANTITRUST IMPACT. 16 Antitrust impact is “the causal link between the antitrust violation and the damages sought 17 by plaintiffs.” The class must provide “some means of determining that each member of the 18 class was in fact injured.” New Motor Vehicles, 522 F.3d at 19 n. 18, 28. Yet the theory of 19 impact must not only be closely tied to the theory of liability, it must be limited to it. See 20 Comcast, 569 U.S. at 35. This follows from the notion that “the defendants cannot be held 21 liable for damages beyond the injury they caused.” See In re Nexium Antitrust Litigation, 777 22 F.3d 9, 18 (1st Cir. 2015). Put simply, a theory of liability (and, thus, class definition) that 23 either catches too many uninjured putative class members, or ignores too many injured ones, 24 likely misses the actual cause of the harm. In sum, then, the class “must be able to establish, 25 predominantly with generalized evidence, that all (or nearly all) members of the class suffered 26 damage as a result of [defendants’] alleged anti-competitive conduct.” In re Qualcomm 27 Antitrust Litigation, 328 F.R.D. 280, 299 (N.D. Cal. 2018). Faced with an expert in each 1 extent” that they weigh on the requirements for certification. See Ellis v. Costco, 657 F.3d 970, 2 981–84 (9th Cir. 2011); Amgen, 568 U.S. at 465–66. 3 Here, the driving question remains “what did defendants’ scheme do to the market?” 4 Asking “which class members were swept up in scheme” poses a secondary question, which, 5 though it often requires individualized inquiry, may still be answered predominantly by 6 common drug sales data. The class offers the testimony of economist Dr. Jeffrey J. Leitzinger, 7 Ph.D., who concludes, based upon the common evidence of economics literature, defendants’ 8 own market forecasts, and both defendants’ and third-party brand and generic Glumetza sales 9 data, that defendants’ anticompetitive conduct inflicted overcharges on each putative class 10 member. 11 Dr. Leitzinger explains that economics literature reliably concludes generic drugs, as 12 essentially-identical commodities, compete with the brand and with each other based on price, 13 lowering prices. Generic drugs crash brand monopolies, entering the market at significant 14 discounts. Studies show this discount increases over time (and as more generics enter the 15 market) and quickly pervades the market as the generics capture the prescription base through 16 permissive or (often, state or locally) mandated generic-substitution at the pharmacy. And, this 17 generic market penetration increases over time. Atop that, some studies also found the brand 18 prices themselves dropped in response to generic competition (Leitzinger Decl., Dkt. No. 246-2, 19 at ¶¶ 19–21, 30–36, 47). See, e.g., New York ex rel. Schneiderman v. Actavis, 787 F.3d 638, 20 644–45 (2d Cir. 2015) (brief survey of state substitution laws). 21 To counter this threat, brands often market an “authorized generic” to recapture some of 22 the market. But, of course, the authorized generic acts like a regular generic, driving greater 23 generic discounts and market capture. So, even though the authorized generic recaptures brand 24 revenue, it still aids competition (id. at ¶¶ 22–24). 25 Our defendants certainly expected as much. Dr. Leitzinger reports, from careful review of 26 their own documents, both Bausch and Lupin expected the coming generics to slash prices and 27 capture nearly the entire Glumetza market after a couple of years. Reality confirmed these 1 at an average discount of 13 percent off the list price. In February 2016, Lupin joined the 2 market at a 47 percent average discount. Bausch’s February 2017 authorized generic then drove 3 the average generic discount to 62 percent and its own brand discount to 18 percent. And, Teva 4 and Sun’s mid-2017 entry drove the generic discount to 93 percent. By 2018, generics had 5 captured 80 percent of the Glumetza prescription base (id. at ¶¶ 37–38, 40). 6 Defendants stalled these discounts by inducing Lupin to delay entry by four years. By 7 February of 2016, the market should have enjoyed years of Lupin’s generic and an authorized 8 generic along with the accompanying (and ever increasing) generic discount and market 9 penetration. Instead, putative class members suffered three and a half years without the 10 discount — and indeed, due to defendants’ further anticompetitive conduct, an 800 percent 11 price hike in 2015. Simply, Dr. Leitzinger concludes, “conduct which delays or limits generic 12 competition . . . inflate[s] prices broadly across the brand prescription base.” More important 13 for the present purposes, however, Dr. Leitzinger also stresses that his model relies only upon 14 the common evidence of, as noted, economic literature, defendants’ own forecasts, and actual 15 sales data (id. at ¶¶ 21, 26–29, 36, 45–51). 16 Dr. Leitzinger finally explains that three forms of overcharge span the class, beginning 17 with lost generic conversion. The more than eighty percent of Glumetza prescriptions that 18 should have been filled with the generic, weren’t. Each of the putative class members served a 19 sufficient population, so each almost certainly lost out on generic substitution of at least one 20 prescription. Every lost generic substitution meant an overcharge. Beneath that, simply due to 21 suppressed competition, each brand and generic purchase also carried an overcharge (id. at ¶¶ 22 54–62). 23 Defendants try to undercut the Dr. Leitzinger’s factual assumptions. Actually, they start 24 by attacking Dr. Leitzinger’s reliance on any factual assumptions at all, charging him with 25 “offer[ing] no evidence for any of his assumptions at all.” But of course Dr. Leitzinger assumes 26 the circumstances in which he models the impact of defendants’ behavior. Recall, antitrust 27 impact is “the causal link between the antitrust violation and the damages.” New Motor 1 impact requires us to assume class counsel will otherwise prove (by common evidence) the 2 antitrust violation at trial. So, Dr. Leitzinger assumes (as counsel directed): (i) Lupin would 3 have entered the market (perhaps even “at risk”) in May 2012; (ii) Santarus would have 4 promptly marketed an authorized generic; and (iii) the 2015 price hikes would not have 5 occurred. 6 In response, defendants primarily contend that all these circumstances must be proven by 7 a preponderance, else they undermine class cohesion. These objections test the merits and run 8 headlong into Amgen. There, the United States Supreme Court explained that in securities fraud 9 the materiality of a misrepresentation — an admittedly essential step in the plaintiffs’ ability to 10 prove classwide reliance on a misrepresentation — did not need to be proven at class 11 certification because, first, materiality turned on an objective (and thus, collective) standard 12 and, more importantly, that materiality comprised an essential element of the case. Individual 13 questions would not predominate absent materiality. Rather, that class would lose. 568 U.S. at 14 466–70. The Supreme Court recently reaffirmed this understanding, explaining that where “the 15 concern about the proposed class is not that it exhibits some fatal dissimilarity but, rather, a 16 fatal similarity — [an alleged] failure of proof as to an element of the plaintiffs’ cause of action 17 — courts should engage that question as a matter of summary judgment, not class certification.” 18 Tyson, 136 S. Ct. at 1047. 19 Dr. Leitzinger’s factual assumptions satisfy both prongs here because each comprises a 20 scenario that the class alleges would have occurred but-for distinct portions of defendants’ 21 anticompetitive conduct. This conduct will be provable by common evidence at trial, as 22 discussed above. More importantly, though, the class’s failure on any ground does not result in 23 rampant individual questions. Instead, that mode of harm drops out of the case. 24 Defendants argue extensively that Lupin would not have launched around May 2012, 25 much less “at risk,” because it expected to lose, and in fact would have lost, the patent suit. 26 Even if it did launch “at risk,” defendants say, Lupin’s generic would have been pulled off the 27 market promptly. In that scenario, defendants’ expert Dr. Bruce Strombom concludes only a 1 weeds we will need to delve at summary judgment or at trial. Actavis itself indicates that “[i]t is 2 normally not necessary to litigate patent validity” — and presumably also infringement — “to 3 answer the antitrust question,” because the lawfulness of the settlement turns on its purpose, 4 effect, and (if needed) justification. See 570 U.S. 157–59. This follows somewhat from the fact 5 that courts remedy analogous anticompetitive behavior, patent misuse, by holding a patent 6 unenforceable, a remedy unnecessary unless the patent is valid and infringed. See Princo Corp. 7 v. Int’l Trade Comm’n, 616 F.3d 1318, 1328 (Fed. Cir. 2010). Then again, given the present 8 analysis turns so heavily on defendants’ intent, some scrutiny of validity and infringement, or at 9 least the boundaries of the parties’ reasonable beliefs about them, may be required. Cf. Alaska 10 Airlines, 948 F.2d at 540; Aspen, 472 U.S. 602–03. But all that will be decided for the class at 11 summary judgment or trial. 12 The point here is that defendants raise the patent merits to change the but-for scenario and 13 jump ahead to the impact analysis, forgetting to reevaluate the threshold question of the antitrust 14 violation. The class’s allegations of an unlawful reverse settlement depend on Lupin’s February 15 2016 market entry being delayed from around mid-2012. If the jury finds Lupin’s generic 16 Glumetza did infringe valid patents, or perhaps at least that both parties honestly believed it did, 17 then negotiations would have started with a late 2016, or even 2020 (after the last patent 18 expired), market entry. In that case, the settlement ceases to be anticompetitive. The class 19 doesn’t just break down — the entire class loses (Aug. 6 Hrg. Tr., Dkt. No. 346, at 15). See 20 Actavis, 570 U.S. 136; Amgen, 568 U.S. at 466–70. So, the patent merits and the assumption 21 that Lupin would have marketed early and successfully do not require proof at this stage. 22 Assuming, then, that the jury does find the 2012 settlement agreement unlawful, it still 23 need not accept all of the allegations. For example, if it finds that Santarus would not have 24 launched an authorized generic anyway, then that circumstance would have applied across the 25 board, to all class members. Moreover, Dr. Leitzinger discusses an authorized generic’s 26 influence on price discounts and generic market penetration as a discrete factor. That means he 27 can easily remove those effects from the model to suit the finding, just as Comcast requires 1 decision not to market an authorized generic would yield rampant individual questions (Dkt. 2 No. 262-5 at 14). 3 So too for the 2015 price hikes. We have excellent counsel on both sides in this case and 4 it will be most engaging to see how counsel argue that the 800 percent price hike served 5 competition. If the argument works, though, then that conduct would not be an antitrust 6 violation, to all class members. And, again, instead of defendants explaining how this would 7 undermine the model, Dr. Leitzinger explains he could easily remove that assumption from it 8 (Leitzinger Decl. at ¶ 70 at n. 86). 9 So, rather than just assuming the class’s cohesion, Dr. Leitzinger simply assumes (as he 10 must) that the class will prove the first portion of its case at trial. If it doesn’t, the harms don’t 11 become individualized. Instead, the class will either have lost the case, or the harms will simply 12 fall away entirely and Dr. Leitzinger will remove them from his model. Defendants’ merits 13 arguments will have to wait for summary judgment or our jury. Thus, this order does not 14 address the class’s evidence of causation submitted in reply and defendants’ objection to 15 consideration of that evidence is denied as moot. 16 Defendants’ invoke In re Lamictal Direct Purchaser Antitrust Litigation, in which the 17 Court of Appeals for the Third Circuit vacated a class certification order for inadequately 18 addressing “a market characterized by individual negotiations and a discounted-brand 19 competition strategy.” That decision, however, turned on the “nuance in the anti-epilepsy drug 20 market.” GlaxoSmithKline, the brand manufacturer, believed an authorized generic would not 21 compete effectively with Teva’s anticipated generic given doctors’ particular reluctance to 22 switch patients’ anti-epilepsy medications. Glaxo thus planned to compete with Teva’s generic 23 based on price in an aggressive individual discount and rebate campaign with pharmacies. Teva 24 learned of this plan and dropped its generic price even lower before launch. These unique 25 pricing facts do not apply here. See 957 F.3d 184, 189–90, 192 (3d Cir. 2020); see also In re 26 Zetia (Ezetimibe) Antitrust Litigation, No. MD 18-02836 DEM, 2020 WL 3446895, *31 (E.D. 27 Va. June 18, 2020) (Magistrate Judge Douglas E. Miller) (distinguishing Lamictal for same C. ESTIMATE OF DAMAGES. 1 Having proven a common mode of antitrust impact, we turn to the damages model. A 2 putative class must “establish that damages are susceptible of measurement across the entire 3 class.” Comcast, 569 U.S. at 35. Yet “[a]ntitrust plaintiffs may satisfy the predominance 4 requirement by using a model that estimates the damages attributable to the antitrust injury, 5 even if more individualized determinations are needed later to allocate damages among class 6 members.” In re Suboxone (Buprenorphine Hydrochlorine and Naloxone) Antitrust Lit., ___ 7 F.3d ___, 2020 WL 4331523, at *4 (3d Cir. July 28, 2020). Indeed, our court of appeals has 8 “repeatedly confirmed” that “the need for individual damages calculations does not, alone, 9 defeat class certification.” Vaquero v. Ashley Furniture Indus., 824 F.3d 1150, 1155 (9th Cir. 10 2016). And as above, “a plaintiff’s damages case must be consistent with its liability case, 11 particularly with respect to the alleged anticompetitive effect of the violation.” Comcast, 569 12 U.S. at 35. 13 Yet “[c]alculations need not be exact.” Ibid. The Supreme Court has repeatedly 14 explained that courts should afford plaintiffs relatively broad leeway in constructing a damages 15 model — at least within the “just and reasonable inference[s] from the proof of defendants’ 16 wrongful acts and their tendency to injure plaintiffs’ business.” Bigelow v. RKO Radio 17 Pictures, 327 U.S. 251, 264 (1946). The Supreme Court has more recently affirmed this 18 understanding, explaining that “[t]he vagaries of the marketplace usually deny us sure 19 knowledge of what plaintiff’s situation would have been in the absence of the defendant’s 20 antitrust violation.” J. Truett Payne Co. v. Chrysler Motors Corp., 451 U.S. 557, 566 (1981). 21 Plainly, we can’t know exactly what would have happened in that but-for world; defendants saw 22 to that. 23 Dr. Leitzinger calculates the aggregate class overcharge, the difference between what 24 class members actually paid and what they should have paid, as a function of the volume of 25 Glumetza (or generic) actually sold, the wholesale acquisition cost, the actual (brand or generic) 26 price, and the brand discount, generic discount, and generic market penetration but for 27 1 defendants’ conduct (Leitzinger Decl. at ¶¶ 63–65). Common evidence drives each of these 2 values. 3 The volume and actual price of Glumetza are easy; Dr. Leitzinger just uses the actual 4 numbers. For the wholesale acquisition cost, Dr. Leitzinger takes the actual value from 2015, 5 just before the exorbitant 800 percent price hikes. And, for generic market penetration, Dr. 6 Leitzinger takes the actual numbers from Lupin’s market capture after it joined in 2016. He just 7 applies that market penetration to 2012, when Lupin was supposed to (allegedly) enter the 8 market. So too with the brand discount. Dr. Leitzinger simply takes the actual observed brand 9 discount after each generic entered the market and, again, backcasts the data to a 2012 Lupin 10 entry (id. at ¶¶ 66–68). 11 Calculating the generic discount takes a little more work, given Lupin joined in the 12 conspiracy, knew it would face no authorized generic (indeed no generic competition at all) for 13 the first year, and almost certainly priced accordingly. Thus, Dr. Leitzinger uses Lupin’s actual 14 discount as evidence, but he also factors in defendants’ forecasts of generic pricing and data 15 from earlier economic studies (id. at ¶ 69). 16 At this stage, Dr. Leitzinger estimates the total class overcharge at approximately $2.3 17 billion (id. at ¶ 73). He also provides some sample calculations, but we need not dig that deep. 18 For now, it suffices that his model employs common data and applies class wide, even if 19 allocation of the damages will require some individualized inquiry down the road. Vaquero, 20 824 F.3d at 1155. Defendants offer three general objections. 21 First, defendants broadly contest various inputs into Dr. Leitzinger’s damages model. To 22 start, we can quibble about the particular inputs for Dr. Leitzinger’s model later. It will be for 23 the jury to determine what exactly happened absent defendants’ conduct. As another judge of 24 this district has observed regarding similar critiques of, coincidentally, Dr. Leitzinger’s earlier 25 work, “[t]h[e]se disputes are not appropriately resolved at this juncture; that the experts dispute 26 what the appropriate inputs should be does not undermine the approach or the reliability of [Dr.] 27 Leitzinger’s model.” See In re Lidoderm Antitrust Litigation, No. MD 14-02521, 2017 WL 1 679367 at *12 (N.D. Cal. Feb. 21, 2017) (Judge William H. Orrick). Yet both of defendants’ 2 arguments on this front also contain their own defects. 3 Defendants contend Dr. Leitzinger’s method of backcasting generic penetration and brand 4 discounts “is unreliable because it fails to recognize the significant differences between the 5 actual world and [Dr.] Leitzinger’s assumed but-for world.” Bigelow and Chrysler undercut 6 this argument. It remains difficult, scratch that, impossible to know what would have happened 7 absent defendants’ unlawful conduct because, assuming the class wins, the unlawful scheme 8 itself deprived us of that information. Estimates are, therefore, the only way to replay the film 9 without the violation. Dr. Leitzinger puts forth a sound enough model and our jury will 10 construct the but-for world, not with absolute precision, but with “just and reasonable 11 estimate[s].” See Bigelow, 327 U.S. at 264–66; Chrysler, 451 U.S. at 565–67. 12 Defendants then object that Dr. Leitzinger’s model improperly includes purchases by opt- 13 out plaintiffs, exclusion of which will require individualized inquiries. Removing opt-out 14 plaintiffs from the calculations may well require individualized inquiries, as most damages 15 inquiries do. But that does not undermine class certification, particularly where the core- 16 classwide model holds. Dr. Leitzinger explains that he can easily adapt his model to account 17 for opt-out plaintiffs. More importantly, however, as Dr. Leitzinger notes, unless the opt-out 18 plaintiffs drop their cases we’re going to be calculating their damages anyway. So the 19 modification to the class damages model will most likely involve subtracting whatever damages 20 we calculate due to each opt-out plaintiff (Leitzinger Rebut. at ¶¶ 47–55). See Vaquero, 824 21 F.3d at 1155. 22 Second, defendants protest that Dr. Leitzinger ignores the post-initial-sale complexities of 23 the pharmaceutical drug market. Both forms of this argument mistake the cognizable harm 24 here, the overcharge. For one, defendants argue that Dr. Leitzinger fails to account for generic 25 bypass. In the but-for world, so the argument goes, purchasers often bypass the wholesaler and 26 buy directly from the generic manufacturer, so defendants’ conduct actually benefitted the 27 wholesaler. Before addressing this objection directly, it’s worth noting that the decision cited to 1 been met with almost universal criticism.” In re Glumetza, No. C 19-05822 WHA, Dkt. No. 2 305, 2020 WL 3498067 at *12 (N.D. Cal. June 29, 2020) (Magistrate Judge Robert M. Illman); 3 see also Braintree Labs. v. McKesson Corp., No. 11-80233 MISC JSW, 2011 WL 5025096 4 (N.D. Cal. Oct. 20, 2011) (Magistrate Judge Jacqueline Scott Corley). More importantly, 5 however, defendants’ argument incorrectly focuses on the wholesaler’s later realized profit 6 margin instead of the actual injury here, the overcharge. Hanover Shoe v. United Shoe Mach. 7 Corp., 392 U.S. 481, 489 (1968). Judge Curtis L. Collier of the Eastern District of Tennessee 8 put it well:

9 Defendant[s] ignore[] the fundamental import of Hanover Shoe and its progeny: a direct purchaser may recover the full 10 amount of the overcharge, even if he is otherwise benefitted, because the antitrust injury occurs and is complete when the 11 defendant[s] sell[] at the illegally high price. That is, the focus of the antitrust laws is limited to the anticompetitive sale. When 12 Defendant[s] sold the [drug] to the wholesalers at an allegedly anticompetitive price, the injury was complete. The jury need not 13 hear any more. True, to calculate damages the jury must consider what the cost of the drug would have been in the absence of an 14 antitrust violation. But Hanover Shoe and Illinois Brick make clear that courts and juries will not be forced down the rabbit hole 15 of hypothetical issues antitrust violators may raise to minimize their liability. 16 In re Skelaxin (Metaxalone) Antitrust Litigation, No. MD 12-02343 CLC, 2014 WL 2002887, 17 *5 (E.D. Tenn. May 15, 2014) (quotations and citations omitted); see also In re Niaspan 18 Antitrust Litigation, No. MD 13-02460 JED, 2015 WL 4197590, *1 (E.D. Pa. July 9, 2015) 19 (Judge Jan E. DuBois) (quoting same). 20 Defendants also forget that the overcharge “arose at the time the extra charge was paid,” 21 regardless of whether the defendant later attempts to mitigate. See Adams v. Mills, 286 U.S. 22 397, 407 (1932). So, their argument that Dr. Leitzinger’s damages model “does not reflect the 23 many discounts and subsequent payments that must be accounted for to calculate the ultimate 24 net price” falls flat. As the Court of Appeals for the First Circuit rejected the same argument in 25 Nexium, “antitrust injury occurs the moment the purchaser incurs an overcharge, whether or not 26 that injury is later offset.” 777 F.3d at 27; see also Hawaii v. Standard Oil of Cal., 405 U.S. 27 251, 262 n. 14 (1972). 1 Third, defendants argue the Dr. Leitzinger cannot adjust his damages model to account for 2 various putative class members what will be barred by the statute of limitations. Assuming they 3 will be barred, Dr. Leitzinger nonetheless explains that those class members’ damages can be 4 easily removed from the calculation of class damages (Leitzinger Rebut. at ¶ 71). Moreover, 5 this objection appears to misunderstand what the limitations period does here. Our court of 6 appeals, when first applying fraudulent concealment to § 15b’s four-year limitations period, 7 rejected the abstract notion that the limitations period was “an inherent element of the right.” 8 The limitations period did not “expunge a cause of action, but rather . . . create[d] a uniform 9 period throughout the entire nation for the commencement of suit.” Simply, a limitations period 10 does not mean that no harm has occurred, it merely marks where claims have grown “stale” and 11 we no longer permit recovery. If defendants’ conduct harmed the market as a whole, and the 12 statute of limitations bars some actors but not others, Dr. Leitzinger’s model should still account 13 for all of the harm to all of the actors. We may, however, simply bar some of those putative 14 class members from recovering if the statute of limitations applies. See Westinghouse Elec. 15 Corp. v. PG&E, 326 F.2d 575, 579–80 (9th Cir. 1964). 16 2. TYPICALITY. 17 Typicality tests “whether other members have the same or similar injury, whether the 18 action is based on conduct which is not unique to the named plaintiffs, and whether other class 19 members have been injured by the same course of conduct.” Wolin v. Jaguar Land Rover N. 20 Am., 617 F.3d 1168, 1175 (9th Cir. 2010). 21 Defendants says the named plaintiffs, assignees of other Glumetza purchasers, are atypical 22 of the class. At the start, though, “an assignee stands in the shoes of his assignor, deriving the 23 same but no greater rights and remedies than the assignor then possessed.” In re Loestrin 24 Fe 24 Antitrust Litigation, No. MD 13-02472 WES, 2019 WL 3214257 at *11 (D.R.I. July 2, 2019) 25 (Judge William E. Smith) (quoting, ultimately, Fox-Greenwald Sheet Metal Co. v. Markowitz 26 Bros., Inc., 452 F.2d 1346, 1358 n. 69 (D.C. Cir. 1971)). For largely the same reasons the Court 27 of Appeals for the Third Circuit has explained, also in the pharmaceutical-antitrust context at 1 right and should be treated as any other class member (absent impropriety in assignment). See 2 In re Modafinil Antitrust Litigation, 837 F.3d 238, 251 (3d Cir. 2016). Assignee status alone 3 here does not weigh against the named plaintiffs. 4 But, defendants contend, the assignments all cover different time periods. Surely putative 5 class members seeking damages from different time periods will have inconsistent goals. 6 Hardly. Regardless of the time period of an assignment or when a putative class member 7 purchased Glumetza, the earlier defendants’ generics entered the market, the earlier generic 8 competition would have lowered prices for all. The earlier the entry, the greater the overcharge, 9 the greater the damages. See Loestrin, 2019 WL 3214257 at *12. 10 Defendants insist, though, that the named plaintiffs didn’t actually buy that much 11 Glumetza compared to, say, the “Big Three wholesalers” who accounted for 82 percent of direct 12 brand purchases and 71 percent of direct generic purchases. “Notably,” defendants say, “none 13 of them has filed suit.” If defendants mean to imply their lawful conduct from this fact, they 14 read too much from the large wholesalers’ absence. Many factors drive a decision not to sue, 15 including, as the Supreme Court has recognized, a wholesaler’s “fear of disrupting relations 16 with their suppliers.” See Illinois Brick, 431 U.S. 720, 746 (1977). More importantly, that the 17 remaining class members possess noticeably smaller recovery interests actually weighs in favor 18 of the economies of class treatment, not against it. See Lidoderm, 2017 WL 679367 at *15. 19 Defendants finish with the argument that the assignments “are complex documents” which 20 assign claims arising from certain portions of sales to different assignees. Tracing the 21 assignments, they say, will be difficult. Maybe so, but no more difficult than we might already 22 expect to calculate to ensure that all damages are no more than “just and reasonable estimate[s]” 23 of the actual harm to each putative class member. See Bigelow, 327 U.S. at 264–66. 24 3. ADEQUACY. 25 Our court of appeals has explained that adequate representatives (1) have no conflicts of 26 interest with other class members, and (2) will prosecute the action vigorously on behalf of the 27 class. Staton v. Boeing Co., 327 F.3d 938, 957 (9th Cir. 2003). 1 Defendants mainly contend that named plaintiffs have not diligently pursued this case and, 2 thus, cannot adequately represent the class. For one example, defendants say that named 3 plaintiffs have not adequately gathered relevant sales information from their wholesaler 4 assignors. They cite one instance in February when wholesaler McKesson ignored a deficient 5 subpoena from named plaintiff KPH, who then never followed up. Vagueness mars defendants’ 6 objection. Without detail sufficient to conclude that the class needs this information to 7 diligently pursue its case, the failure to follow up on a discovery request means little. 8 For the other, defendants object to two instances of named plaintiffs’ supplemental 9 document production. Parties must conduct a reasonably comprehensive document search, not 10 an absolutely comprehensive one. See, e.g., SUPPLEMENTAL ORDER TO ORDER SETTING INITIAL 11 CASE MANAGEMENT CONFERENCE ¶ 16. Relevant documents often appear in later, different 12 searches and nonetheless require prompt, supplemental disclosure. Defendants neither contend 13 that the named plaintiffs withheld documents nor that they conducted inadequate initial 14 searches. Nothing here raised undercuts named plaintiffs’ adequacy to represent the putative 15 class. 16 Defendants also argue that named plaintiffs sue over their substantial assigned brand 17 Glumetza purchases, but most of the class purchased only generic Glumetza. So, while named 18 plaintiffs seek damages from actual sale overcharges, absent class members might prefer a 19 damages model based primarily on all the generic purchases and resales they never made due to 20 defendants’ conduct. Maybe so. But our court of appeals is not alone in warning against denial 21 of class certification on the basis of speculative conflicts. See In re Online DVD-Rental 22 Antitrust Litigation, 779 F.3d 934, 942 (9th Cir. 2015); Kohen v. Pacific Inv. Mgmt. Co., 571 23 F.3d 672, 680 (7th Cir. 2009). Should any absent class members prefer a different damages 24 model, they remain free to opt-out of the class. In re Niaspan Antitrust Litigation, 397 F. Supp. 25 3d 668, 681 (E.D. Pa. 2019); Lidoderm, 2017 WL 679367 at *15. 26 4. SUPERIORITY. 27 Even if common questions predominate, a Rule 23(b)(3) class must be the superior 1 (A) the class members’ interests in individually controlling the prosecution or defense of separate actions; 2 (B) the extent and nature of any litigation concerning the 3 controversy already begun by or against class members;

4 (C) the desirability or undesirability of concentrating the litigation of the claims in the particular forum; and 5 (D) the likely difficulties in managing a class action. 6 Though complexities will surely arise, including some individualized questions, no 7 inability to manage a class action manifests here. The finality of decision for all involved 8 counsels in favor of concentrating this litigation in a single forum — indeed the class members 9 came to that conclusion themselves by all filing here. No other cases arising from defendants’ 10 challenged conduct here remain before any other tribunal. All this, in conjunction with the 11 predominance of common issues, establishes that a class action is the superior method of 12 adjudicating this case. 13 It’s also worth noting the simplicity of the proposed class definition: Glumetza purchasers 14 from defendants since May 2012. If, given defendants’ litany of arguments, more specific 15 requirements become necessary, the class definition will remain manageable. So too, if specific 16 populations appear within the class, subclasses may be certified to proceed. 17 Defendants contend that a class is not superior because the largest purchasers of Glumetza 18 have elected to go it alone. As above, a larger number of smaller claims lend themselves to the 19 economies of the class. Nevertheless, defendants also argue that class members seek, on 20 average, treble damages of $28 million. So, their argument goes, the recovery will cover any 21 legal fees and each putative class member can proceed individually with its own counsel, each, 22 naturally, siphoning the potential recovery as though we’re trying to recreate Dickens’ Bleak 23 House. A class remains superior here. 24 5. NUMEROSITY. 25 A class must be “so numerous that joinder of all members is impracticable.” Rule 26 23(a)(1). “The numerosity requirement requires examination of the specific facts of each case 27 and imposes no absolute limitations.” Gen. Tel. Co. of the Northwest, Inc. v. EEOC, 446 U.S. 1 318, 330 (1980). Defendants do not contest the numerosity of the eighty-one member proposed 2 class. 3 6. ASCERTAINABILITY. 4 Finally, though not an explicit component of Rule 23, and one to which our court of 5 appeals has not yet definitively spoken, many courts in this circuit have long imposed the 6 requirement that “[a] class definition should be precise, objective, and presently ascertainable.” 7 See Senne v. Kansas City Royals Baseball Corp., 315 F.R.D. 523, 542 (N.D. Cal. 2016) (Chief 8 Magistrate Judge Joseph C. Spero); Xavier v. Philip Morris USA, 787 F. Supp. 2d 1075, 1089– 9 90 (N.D. Cal. 2011); O’Connor v. Boeing N.A., 184 F.R.D. 311, 319 (C.D. Cal. 1998) (Judge 10 Audrey B. Collins). Among others, the requirement “allow[s] potential class members to 11 identify themselves for purposes of opting out of a class.” Carrera v. Bayer Corp., 727 F.3d 12 300, 307 (3d Cir. 2013). 13 Though the proposed class definition turns on objective criteria, dates and purchases, it 14 remains an open class. That is, by seeking to include within the class purchasers between May 15 2012 “until the effects of the defendants’ conduct ceased,” the class by definition includes those 16 who have not yet purchased Glumetza from a defendant. Obviously, until such purchase 17 occurs, those purchasers remain unknown. And, until they become known, we cannot notify 18 them of this class and honor their decision to, if they so choose, opt-out. Thus, the class 19 definition must be limited to those to whom we may provide effective notice at this time. 20 CONCLUSION 21 For the foregoing reasons, the following class is CERTIFIED:

22 All persons or entities in the United States and its territories who directly purchased Glumetza or generic Glumetza from a defendant 23 from May 6, 2012 until the date of this order. 24 Shana Scarlett and Lauren Barnes of Hagens Berman Sobol Shapiro LLP, Steve Shadowen of 25 Hilliard & Shadowen LLC, and Joseph Vanek of Sperling & Slater, P.C. have represented the 26 class effectively since their appointment as interim counsel (Dkt. No. 70) and are CONFIRMED 27 as co-lead counsel for the class. By AUGUST 31, 2020 AT NOON, the parties shall jointly submit 1 mail and internet notice. In crafting their joint proposal, the parties shall please keep in mind 2 the guidelines for notice to class members given in the NOTICE AND ORDER RE PUTATIVE CLASS 3 ACTIONS AND FACTORS TO BE EVALUATED FOR ANY PROPOSED CLASS SETTLEMENT (Dkt. No. 4 39). 5 IT IS SO ORDERED. 6 7 Dated: August 15, 2020. 8 9 Al = ] | . 4 WILLIAM ALSUP 10 UNITED STATES DISTRICT JUDGE 11 12

Z 18 19 20 21 22 23 24 25 26 27 28

In Re Glumetza Antitrust Litigation, (N.D. Cal. 2020).

Pincite

In Re Glumetza Antitrust Litigation (In Re Glumetza Antitrust Litigation) — published by Counsel Stack Legal Research, free access to 12M+ legal documents.