U.S. v. Premo Pharmaceutical Laboratories

Full title: UNITED STATES of America, Plaintiff, v. PREMO PHARMACEUTICAL LABORATORIES…

Court: United States District Court, D. New Jersey

Date published: Jan 20, 1981

511 F. Supp. 958 (D.N.J. 1981)

Opinion

Civ. No. 80-699.

January 20, 1981.

Charles J. Walsh, Asst. U.S. Atty., Newark, N.J., Donald O. Beers, Associate Chief Counsel for Enforcement of Food and Drug Administration, Rockville, Md., for plaintiff.

Henry J. Catenacci, Podvey Sachs, Newark, N.J., John S.E.ldred, and Deborah Shur Trinker, Keller Heckman, Washington, D.C., for defendants.

OPINION

LACEY, District Judge.

INTRODUCTION

The United States sues to enjoin defendants from continuing their alleged violations of the Federal Food, Drug, and Cosmetic Act (Act), 21 U.S.C. § 301-392 (1976 Supp. I 1977). These violations, as charged in the complaint, are that defendants are marketing human drugs in interstate commerce without the requisite approval of the Food and Drug Administration (FDA), and without the testing required to show that the drugs are safe and effective for their intended uses.

Defendant Premo Pharmaceutical Laboratories, Inc. (Premo) is a manufacturer and distributor of prescription drugs. Defendant Seymour N. Blackman, Premo's president, has responsibility for and authority over operation of the corporation. Involved are the following drug products manufactured by Premo: (1) triamterene with hydrochlorothiazide capsules, marketed by Premo under the name Triamthiazide — a generic version of Dyazide, manufactured by Smith, Kline and French (used in the treatment of hypertension and edema); (2) allopurinol tablets — a generic version of Zyloprim, marketed by Burroughs Wellcome (used in the treatment of the symptoms of gout); (3) chlorthalidone tablets — a generic version of the chlorthalidone tablets manufactured by USV Pharmaceuticals under the name Hygroton (used in the treatment of hypertension and edema); (4) betamethasone valerate cream, marketed by Premo under the name Beta Val — a generic version of Valisone, produced by Schering Company (used in the treatment of various dermatological conditions); (5) trifluoperazine hydrochloride tablets — a generic version of the trifluoperazine tablets marketed by Smith, Kline and French under the name Stelazine (used in the management of manifestations of psychotic disorders); (6) doxylamine succinate and pyridoxine hydrochloride tablets, marketed by Premo under the name Doxine — a generic version of Bendectin, manufactured by Merrell-National Laboratories (used to relieve nausea and vomiting during the early months of pregnancy); (7) hydroxyzine pamoate capsules, marketed by Premo under the name Hy Pam — a generic version of the antihistamine Vistaril produced by Pfizer Pharmaceuticals (used in the management of emotional distress, anxiety, tension, and psychomotor agitation); and (8) hydroxyzine hydrochloride tablets — a generic version of the antihistamine Atarax produced by Pfizer (uses similar to that for Vistaril). As indicated, these products are all generic (or "me-too") versions of FDA-approved "brand name," or "pioneer," medications. It is Premo's version of each of these established products that is in dispute here.

Specifically, the Government charges that the eight drugs just described are "new drugs" within the meaning of 21 U.S.C. § 321(p) (1976), and thus Premo is in violation of section 355(a) of the Act for shipping these products without having first submitted to FDA, and having had approved, a new-drug application (NDA) or an abbreviated new-drug application (ANDA). Section 321(p) provides:

The term "new drug" means —

(1) Any drug (except a new animal drug or an animal feed bearing or containing a new animal drug) the composition of which is such that such drug is not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof, except that such a drug not so recognized shall not be deemed to be a "new drug" if at any time prior to June 25, 1938 it was subject to the Food and Drugs Act of June 30, 1906, as amended, and if at such time its labeling contained the same representations concerning the conditions of its use; or

(2) Any drug (except a new animal drug or an animal feed bearing or containing a new animal drug) the composition of which is such that such drug, as a result of investigations to determine its safety and effectiveness for use under such conditions, has become so recognized, but which has not, otherwise than in such investigations, been used to a material extent or for a material time under such conditions.

21 U.S.C. § 321(p) (1976). Section 355(a) provides:

No person shall introduce or deliver for introduction into interstate commerce any new drug, unless an approval of an [NDA] is effective with respect to such drug.

Id. § 355(a).

Defendants concede new drugs must receive an approved NDA from FDA prior to marketing; however, they deny that their products are "new drugs" within the meaning of the Act. Defendants' position is based on the fact that the active ingredients — and in some cases the inactive ingredients as well — in their generic medications are the same as those in their FDA-approved, pioneer counterparts.

The United States contends that differences in inactive ingredients between two products with the same active ingredients can materially affect the safety and effectiveness of the medication; that the "same" active ingredient from different sources may cause different effects; and that, even if all ingredients are the same, differences in manufacturing practices can cause one manufacturer's product to differ substantially from another in safety and effectiveness.

At the outset of this action, the Government sought a temporary restraining order. That application, based upon a relatively modest submission, was denied, in part because of defendants' agreement to refrain from further shipments of Triamthiazide (triamterene with hydrochlorothiazide), the medication apparently considered by the United States to have the greatest potential for harm. Thereafter, the Government brought on this application for a preliminary injunction.

Given the extent of the disagreement between the parties, the importance of this issue to the public health and safety, and the uncertain state of the law in this circuit, I afforded the parties a broad opportunity to present their evidence and arguments. The resultant record is perhaps the most detailed one developed to date in any case considering the generic/new-drug issue.¹ Testimony was presented over eleven days on each of the eight products named in the complaint, producing in excess of 2,000 transcript pages. This was supplemented by 2,000 pages of affidavits and depositions, more than 400 pages of pre- and posthearing briefing, and 1,000 pages of other posthearing submissions.

1 Thus, for example, I permitted defendants to introduce, conditionally, secret data developed by Premo regarding the safety and effectiveness of its products, notwithstanding the Government's timely objection that this evidence should be excluded as irrelevant under the Act because general expert recognition of safety and effectiveness can only be based on scientific material that is generally available. The Government was permitted to offer, also conditionally, contradicting evidence. For the reasons set forth below, I now rule that this evidence is, indeed, irrelevant. Nonetheless, a full record is available for appellate review.

For the reasons hereinafter set forth, the Government's motion for preliminary injunctive relief is granted.

A word about the format of this opinion, and my findings and conclusions of law, is appropriate.

The first portion of this opinion, "DISCUSSION," pp. 962-977 infra, is divided into six parts. Part I provides a brief history of federal regulation of the drug industry. Part II discusses factors that can affect the bioavailability and bioequivalence of drug products, two concepts central to this litigation. Parts III and IV examine the scant but developing law on the generic/new-drug issue and analyze the evidence in light of my determination that the law requires me to resolve only this question in deciding whether a drug product is a "new drug" under the Act: Is the product generally recognized as safe and effective by experts qualified by scientific training and experience to make such a determination? Part V considers certain additional arguments raised by defendants, and Part VI concludes that the Government is entitled to injunctive relief.

The second portion of this opinion is intended to advance appellate review in the event the court of appeals decides that the determinative issues are broader than the issue of expert recognition of safety and efficacy. Thus, under my Findings of Fact I have referred to not only the evidence considered under "DISCUSSION," but also that which was received conditionally and is now determined by me to be irrelevant. See p. 961, note I supra.

DISCUSSION I

The Food and Drug Act of 1906, ch. 3915, 34 Stat. 768, was the first legislation of national scope directed at the regulation of drug products. The Act set standards of purity for drugs sold in the United States and required accurate labeling of the drugs' contents. Id. §§ 1, 2, 7. The Act also made unlawful the marketing of adulterated or misbranded drugs and provided for removal of such drugs from the market through libel actions. Id. § 2. There were, however, no provisions regulating false claims of efficacy until the Food and Drug Act of 1906 was amended in 1912 to declare misbranded drugs that were not effective for use under the conditions for which they were recommended.^1a

The greatest defect in the Act, however, was its failure to provide any mechanism for premarketing agency clearance. It was impossible to prevent an unsafe or ineffective drug from reaching the market. In 1938 the "wonder drug" "Elixir of Sulfanilamide," a solution based on diethylene glycol and (significantly for the present case) a presumed harmless, inert solvent ingredient, went on the market. Apparently, no tests for toxicity were performed prior to marketing; and almost one hundred people died before the drug could be withdrawn.

This Sulfanilamide tragedy led to the Federal Food, Drug, and Cosmetic Act of 1938, 21 U.S.C. § 301-392 (1976 Supp. I 1977) (amended 1962), with provisions for premarketing review of new drugs. This review, however, was directed solely to assuring drug-product safety. It was not until the Act was amended in 1962 that the definition of "new drug" was enlarged to include drugs not generally recognized as safe and effective.

The 1962 amendments changed the data-reporting requirements of the "new drug" procedure to require submission of data showing efficacy and, in place of automatic approval of NDAs not disapproved, the procedure under the 1938 Act, the 1962 amendments required positive agency approval to make an NDA effective. See Note, Drug Efficacy and the 1962 Drug Amendments, 60 Geo.L.J. 185 (1971). For further discussion of the history of drug regulation by the federal government, see Weinberger v. Hynson, Westcott Dunning, Inc., 412 U.S. 609, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973); USV Pharmaceutical Corp. v. Weinberger, 412 U.S. 655, 93 S.Ct. 2498, 37 L.Ed.2d 244 (1973); United States v. Articles of Drug (Lannett), 585 F.2d 575 (3d Cir. 1978); Hoffmann-LaRoche, Inc. v. Weinberger, 425 F. Supp. 890 (D.D.C. 1975).

Where a drug product is claimed by its manufacturer to be a copy of a product already approved by FDA on the basis of an NDA, the manufacturer may file with FDA an ANDA which relies upon the safety and effectiveness tests performed with respect to the FDA-approved pioneer product. However, FDA will approve an ANDA only where the "me-too" product is shown to be the therapeutic equivalent of the pioneer and safe and effective in accordance with 21 U.S.C. § 355(d) (1976). See Premo Pharmaceutical Laboratories, Inc. v. United States, 629 F.2d 795, at 798 (2d Cir. 1980) (" Premo "). See generally Hoffmann-LaRoche, Inc. v. Weinberger, supra. No ANDA has been approved for the eight Premo products in suit.

II

Much of the testimony on this hearing related to bioavailability. Bioavailability is "the rate and extent to which the active drug ingredient or therapeutic moiety is absorbed from a drug product and becomes available at the site of drug action." 21 C.F.R. § 320.1(a) (1980). See Dorland's Illustrated Medical Dictionary 200 (25th ed. 1974). If there is no significant difference between the rate and extent of absorption of two drugs administered at the same molar dose of therapeutic moiety under similar experimental conditions, the drugs are said to be bioequivalent. See 21 C.F.R. § 320.1(e) (1980). When two different drug products are to be used interchangeably in the treatment of illness, it can be critical that the products are bioequivalent — that is, that there be no significant difference in the products' bioavailability. A drug that is less bioavailable than that for which it is substituted will deliver less of its active ingredient than expected; a drug that is more bioavailable than that which it replaces presents the danger of overdosage.² See P. Ex. 4, ¶¶ 8-17 (Crout Affidavit); P. Ex. 3, ¶¶ 18-22 (Cortell Affidavit); P. Ex. 6, ¶¶ 9-11 (Perel Affidavit); P. Ex. 2, ¶ 39 (Cabana Affidavit).

2 For example, although Premo insists that its triamterene with hydrochlorothiazide capsule, Triamthiazide, is therapeutically equivalent to Dyazide, Smith, Kline and French's FDA-approved formulation of triamterene with hydrochlorothiazide, there is no indication Premo has determined the precise bioavailability of its product. Premo labels its Triamthiazide as a diuretic-antihypertensive agent for use in treatment of edema (fluid retention in body tissue) associated with the following conditions: congestive heart failure, during which the blood pools in other organs (lungs and liver), because the heart does not perform efficiently in pumping the blood; hepatic cirrhosis, in which destruction of normal liver cells with scarring causes obstruction of circulation that results in the abdominal area filling with retained fluids; the nephrotic syndrome, a kidney disorder, the major symptoms of which are swelling, protein in the urine, and a low level of albumin in the blood; tissue-fluid retention due to the action of drugs such as corticosteroids and estrogens. See P. Ex. 4, ¶ 12 (Crout Affidavit). It is also labeled for use in hypertension, which affects the arteries and heart and can result in heart failure or stroke. Id. If Triamthiazide is substituted by a physician for FDA-approved Dyazide in the belief the two are bioequivalent, and Triamthiazide is actually less bioavailable than Dyazide, then insufficient amounts of the active ingredient will be released into the blood and serious complications, even death, could result. Id. at ¶ 14.
On the other hand, assuming Triamthiazide to be more bioavailable than Dyazide, if a physician has calculated a Dyazide titration for a given patient — carefully monitoring the patient's reaction to the medication, raising the amount administered to achieve the desired result and lowering that amount when signs of adverse reaction are observed — and a patient is switched from Dyazide to Triamthiazide without the physician's knowledge, a triamterene overdosage could result. An effective overdosage of triamterene can produce hyperkalemia, an abnormally high concentration of potassium in the blood. Hyperkalemia, particularly in patients who have impaired kidney function due to diabetes or hypertension or who eat high potassium foods or take potassium containing dietary supplements while on triamterene, is commonly fatal. Id. at ¶ 17. See also Findings of Fact 129-135.

The Government argues that physicians and patients generally believe or assume that generic versions of "brand-name" products are bioequivalent to the respective pioneer drugs themselves. Because the generic versions are usually significantly less expensive than the pioneer drugs, they are frequently freely substituted for their more expensive counterparts. Thus, the Government contends, it is crucial that the products be, in fact, bioequivalent. Contending that, because so many factors independent of a product's active ingredient — and even, as indicated below, of its inactive ingredients — can affect a drug product's bioavailability (and thus its bioequivalence to an FDA-approved medication), the Government concludes that absent preclearance of each of Premo's products named in the complaint, there can be no assurance of its safety and effectiveness, and it must be considered a "new drug" within the meaning of section 321(p) of the Act.

Premo argues that so long as the active ingredient in its product is the same as the active ingredient in the approved pioneer drug all other questions are irrelevant because the active ingredient, the "drug," has been recognized as safe and effective by the FDA.

There are many factors that can affect a drug's bioavailability. Among these are the particle size and crystalline form of the active ingredient; the choice of inactive ingredients (excipients), such as binders and fillers; the facilities and controls used in the manufacture and processing of the drug; and the environmental conditions during manufacture and storage.

The particle size and crystalline form of the active ingredient can be crucial because these factors can affect the drug's solubility and/or the rate that the drug dissolves in gastrointestinal fluids. This, in turn, can affect the bioavailability of the active ingredient itself.³ P. Ex. 1, ¶ 10 (Barr Affidavit); P. Ex. 7, at 4 (Shangraw Affidavit); P. Ex. 2, ¶ 24 (Cabana Affidavit); P. Ex. 4, ¶ 11 (Crout Affidavit).

3 Active ingredients may exist in a number of crystal forms. Although one form (polymorph) is chemically indistinguishable from another, polymorphic forms differ substantially with respect to properties such as density, melting point, solubility, and the rate at which the active ingredient will go into solution. Some polymorphic forms are not stable and after a period of time will go into a more stable, and different, polymorphic form. The drug chloroamphenicol plainly demonstrates the influence of polymorphism on bioavailability. When a particular polymorph is used to produce a chloroamphenicol drug product, the product will produce peak blood concentrations of 3 mcg chloroamphenicol/mg. When a different polymorph is employed exclusively in the product, the same dose will produce peak blood concentrations of 22 mcg/mg. See P. Ex. 7, at 4-5 (Shangraw Affidavit).
Varying the size of the particles of the active ingredient will alter the amount of the article that is exposed to water and thereby change the rate and extent of dissolution of the active ingredient. Id. at 6.

Factors that markedly alter availability include size of crystal or particle, crystalline state of drug since differences in free energy of polymorphs influence solubility, dissolution rate, compression used in tabulation, form of drug (solution or solid), pKa of compound, excipient, presence of material in the same dosage that may complex or react with drug in the physiological environment, chemical content and purity, coating, and sustained release forms.

Calesnick, Intimate Study of Drug Action IV, in Drills Pharmacology and Medicine 113 (4th ed. J. DiPalma 1971). See Fingl Woodbury, Introduction, in The Pharmacological Basis of Therapeutics 8 (5th ed. L. Goodman A. Gilman 1975); M. Rowland, Drug Administration and Regimens, in Clinical Pharmacology 32 (K. Melmon H. Morrelli eds. 1978).
Other structural factors can be significant as well. For example, changing the solvent which is used to prepare the active ingredient or the form of the active ingredient — acid, base, or salt — that is used to formulate the final dosage may result in a change in the properties of the active ingredient. P. Ex. 7, at 5-6 (Shangraw Affidavit).

The choice of inactive ingredients is of great importance, because these ingredients may interact with the product's active ingredient and thereby significantly increase or decrease the bioavailability of the active ingredient. P. Ex. 1, ¶ 10 (Barr Affidavit); P. Ex. 7, at 6 (Shangraw Affidavit); P. Ex. 2, ¶¶ 14, 25 (Cabana Affidavit); P. Ex. 4, ¶ 11 (Crout Affidavit). Diluents, used to provide tablet bulk, can affect bioavailability in numerous ways, such as by interfering with the drug's absorption. P. Ex. 7, at 7 (Shangraw Affidavit). Binders, used to provide tablet cohesion, may retard the rate and extent of dissolution when used inappropriately. Id. And disintegrants and surfactants, added to tablets or capsules to cause them to break up after ingestion, can also affect the drug's dissolution and thus its bioavailability. Id. at 7-8.

Even if the characteristics of a product's active ingredient are kept constant, and precisely the same excipients are used, the pharmacological properties of a "me-too" drug product may still differ substantially from those of the pioneer drug because of the manufacturing process itself. P. Ex. 1, ¶ 10 (Barr Affidavit); P. Ex. 7, at 9 (Shangraw Affidavit); P. Ex. 2, ¶ 24 (Cabana Affidavit); P. Ex. 4, ¶ 11 (Crout Affidavit). In fact, as set forth in an affidavit submitted on behalf of the Government, Dr. Ralph Shangraw, Chairman of the Department of Pharmaceutics at the University of Maryland School of Pharmacy, states:

[E]ven if different manufacturers of the same tablet form drug product start with the same active ingredient (manufactured under the same conditions by the same firm), use exactly the same inactive ingredients from the same sources, and manufacture the product under exactly the same conditions, using exactly the same equipment, it still may be possible that the finished dosage form (tablet) will not perform in an identical manner in the same patient.

P. Ex. 7, at 9. See P. Ex. 1, ¶ 10 (Barr Affidavit); P. Ex. 2, ¶ 21 (Cabana Affidavit); P. Ex. 4, ¶ 11 (Crout Affidavit). Dr. Shangraw also noted that the method by which disintegrants are incorporated into the tablet or capsule may significantly affect the effectiveness of the disintegrant. If an incorrect method of disintegration is employed, the drug product may not disintegrate properly, causing an insufficient quantity of the drug to be absorbed and thereby reducing its effectiveness.

In addition, a multitude of other factors can affect bioavailability. These include: the method of granulation and compression; granule hardness, size, and size distribution; flow properties of the powder mixture; capsule or tablet coating method; mixing time; compression pressure and compression dwell time; composition and concentration of disintegrants, lubricants, and other excipients, and the drug-to-excipient ratio; and even the order of mixing a drug product's various excipients. P. Ex. 7, at 9-16 (Shangraw Affidavit). Finally, Dr. Shangraw described the choice of lubricants — the substances used to reduce adhesion of the drug product, prior to compression, to the punches and dies in the tableting machine — as

the single most significant variable that may affect the bioavailability of a drug product. A change in the type or amount of lubricant used, in the method by which a lubricant is incorporated, or in the length of the blending time may significantly affect the bioavailability of a drug product. Because these substances are hydrophobic (water-repellent), use of an excessive amount of a lubricant will have a "raincoat" effect and will reduce the rate and/or extent of dissolution and, consequently, the amount of drug that will be absorbed by the body.

Id. at 8.

Bioavailability can also be affected by environmental factors. Humidity during manufacture can cause problems ranging from simple discoloration to subpotency. Id. at 17.

See also Findings of Fact 21-51.

III

While many courts have grappled with the question of what constitutes a "new drug" under the Act, see, e.g., United States v. X-Otag Plus Tablets, 602 F.2d 1387 (10th Cir. 1979); United States v. An Article of Drug . . . "Entrol-C Medicated," 513 F.2d 1127 (9th Cir. 1975); USV Pharmaceutical Corp. v. Richardson, 461 F.2d 223 (4th Cir. 1972), aff'd sub nom. USV Pharmaceutical Corp. v. Weinberger, 412 U.S. 655, 93 S.Ct. 2498, 37 L.Ed.2d 244 (1973); United States v. 41 Cases, More or Less, 420 F.2d 1126 (5th Cir. 1970); Tyler Pharmacal Distributors, Inc. v. HEW, 408 F.2d 95 (7th Cir. 1969); United States v. Articles of Drug (Hormonin), 498 F. Supp. 424 (D.N.J. 1980); United States v. Articles of Food and Drug, 444 F. Supp. 266 (E.D. Wis. 1978); United States v. Articles of Drug . . . "Colchicine," 442 F. Supp. 1236 (S.D.N.Y. 1978); United States v. Lanpar Co., 293 F. Supp. 147 (N.D.Tex. 1968), only recently has the question been presented in the context of whether drug products that are generic versions of products with approved NDAs are "new drugs" for the purposes of section 321(p). Prior to this action only a limited number of cases have addressed this issue. In addition to the Premo decisions in the court of appeals and district court [ 475 F. Supp. 52 (S.D.N.Y. 1979)], see United States v. Articles of Drug (Lannett), 585 F.2d 575 (3d Cir. 1978) ("Lannett"); United States v. Generix Drug Corp., 498 F. Supp. 288 (S.D.Fla. 1980); United States v. Pharmacal, Inc., No. 78-3685 (S.D.Fla. Mar. 2, 1979); Pharmadyne Laboratories, Inc. v. Kennedy, 466 F. Supp. 100 (D.N.J.), aff'd on other grounds, 596 F.2d 568 (3d Cir. 1979) (" Pharmadyne").

These decisions delineate essentially three positions on the generic/new-drug issue: (1) that espoused in dictum by the Court of Appeals for the Third Circuit in Lannett, which, broadly interpreted, suggests that a product is not a "new drug" if "its therapeutically active ingredients are identical with those of a recognized and approved drug both chemically and quantitatively";⁴ (2) that formulated by the district court in Premo — and adopted substantially by the district court in Generix — which held that in certain circumstances the question of safety and efficacy, as determinative of the new-drug issue, is one for the court; and (3) that set forth by the Court of Appeals for the Second Circuit in the Premo appeal, reversing the district court, which held that the only issue to be decided by a district court presented with a new-drug question is whether, based on use to a material extent or for a material time and on published studies or other publicly available data, an unapproved product is generally recognized by experts as safe and effective.⁵

4 Pharmadyne Labs, Inc. v. Kennedy, 466 F. Supp. 100, 103 (D.N.J.), aff'd on other grounds, 596 F.2d 568 (3d Cir. 1979). This is the broadest interpretation to which the Lannett dictum is susceptible, and I set it forth here to exemplify one end of the spectrum of judicial opinion in this area. Lannett is also amenable to a substantially narrower and arguably more accurate reading. The court employs language throughout the opinion that suggests there was an identity of ingredients, active and inactive, between the generic and pioneer products at issue in the case. The court refers to the Government's concession that the pioneer and generic products are "the same generically," 585 F.2d at 582, "identical," id. at 584, and "admittedly the same," id. at 584 n. 16. And the court's discussion focuses on the significance, or lack of it, of "each manufacturer's particular production techniques," id. at 583, and "individual manufacturing idiosyncracies," id. at 583 n. 15. Under this interpretation, the only issue addressed in the court's dictum is whether an unapproved product chemically and quantitatively identical to its pioneer counterpart is a "new drug" for purposes of the Act — the question which the Second Circuit specifically left unresolved in Premo, and which is directly presented by at least one of defendants' products here. See pp. 975-976 infra. This interpretation of Lannett is consistent with that reached by a subsequent panel of the Court of Appeals for the Third Circuit on the Pharmadyne appeal. The court stated:

Pharmadyne contends that . . . this court, in a previous decision ( Lannett), has decided the only legal and factual issues involved in the seizure action which are pertinent to this case. The flaw in this argument, however, is that the factual and legal issues presented by Pharmadyne's actions are not clearly identical to those which we decided in Lannett. Here, the district court observed that the FDA adduced evidence that the inactive — and perhaps the active — ingredients used in Pharmadyne's formulation are not identical to those used in the approved drug; and that these differences in the total formulation could create a health risk. . . . We are satisfied that the legal and factual issues raised by this case were not decided in Lannett. . . .

Pharmadyne Labs, Inc. v. Kennedy, 596 F.2d 568, 571 n. 6 (3d Cir. 1979).
On the other hand, elsewhere in the Lannett opinion the court describes Lannett's products as "almost identical" to their pioneer counterparts, 585 F.2d at 582, indicating that a broad interpretation of the case may be the more appropriate after all. As will become apparent in text infra, however, I find the merits of neither interpretation supported by the plain language of the Act, its legislative history, or the evidence developed before me in this case.

5 To the same effect is Pharmadyne Labs, Inc. v. Kennedy, 466 F. Supp. 100 (D.N.J.), aff'd on other grounds, 596 F.2d 568 (3d Cir. 1979), discussed in text infra.

In Lannett the Government brought an action to condemn certain drug products manufactured by the Lannett Company. Lannett's products were generic versions of established pioneer drugs that had been approved by FDA. FDA had labeled them "new drugs" under the Act and had required Lannett to submit an ANDA to the agency for each of the drugs. Prior to 1975, because the pioneer counterparts to Lannett's products had received FDA approval, FDA had permitted Lannett to market its products pending the outcome of its ANDAs. In 1975, however, the United States District Court for the District of Columbia in Hoffmann-LaRoche, Inc. v. Weinberger, 425 F. Supp. 890 (D.D.C. 1975), held that once the FDA denominates a drug product a "new drug," the product may not be marketed without an approved new-drug application; and permitting new drugs to be marketed prior to such approval "contravenes the clear statutory requirement of preclearance mandated by 21 U.S.C. § 355 (1970)." 425 F. Supp. at 894.⁶ Following Hoffmann-LaRoche, FDA altered its marketing policy and ordered a freeze on the marketing of all "new drugs," including those produced by Lannett, until they were approved by FDA. Lannett continued to market its products, and after ordering Lannett to cease marketing, FDA seized the drugs, and the Government brought the action for condemnation. The district court granted summary judgment in favor of the Government.

6 As in Lannett and here, the drug products at issue in Hoffmann-LaRoche were generic versions of approved, pioneer products.

On appeal the Court of Appeals for the Third Circuit reversed on the narrow ground that it was "error for the district court to have precluded Lannett from challenging `new drug' status." United States v. Articles of Drug (Lannett), supra, 585 F.2d at 581. In an effort to "lend guidance to prevent another appeal," id. at 582, however, the court proceeded to discuss, in dictum, the question whether generic versions of approved drugs should be considered "new drugs" under the Act.

The court rejected FDA's position that section 321(p) requires general recognition of safety and effectiveness for each specific drug product and, instead, adopted that espoused by Lannett: Drugs that are "the same generically" as a drug already approved as safe and effective by the FDA are not "new drugs"; section 321(p) requires only general recognition of safety and effectiveness, and "bioavailability, bioequivalence, and other quality control tests are irrelevant as to safety and efficacy." Id. In reaching this conclusion, the court examined the "definitions and interpretations" section of FDA's regulations and — noting that although they do not purport to be exhaustive the regulations make no specific reference to quality control or "a specific product's capabilities" — concluded:

The absence of any regulatory provision setting forth a manufacturer's production techniques as an element making a formulation "new" is powerful evidence that FDA's theory is not based strongly on the statute.

Id. at 583. As a second, related point, the court suggested that, if FDA's position were supported by the Act, section 321(p)(1)

would be superfluous for under the FDA's theory any new manufacture of a generic drug, whether approved or not, would be a "new drug" in the absence of general recognition of the safety and effectiveness of the new manufacture.

Id.

Finally, the Lannett court found support for its position in Weinberger v. Hynson, Westeott Dunning, Inc., 412 U.S. 609, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973) ("Hynson") and United States v. An Article of Drug . . . "Mykocert," 345 F. Supp. 571 (N.D.Ill. 1972) ("Mykocert"). In Hynson the Supreme Court held that FDA withdrawal of a pioneer product's NDA is sufficient grounds for the FDA to force withdrawal of every generic version of the pioneer, provided all manufacturers have had an opportunity to be heard. The court of appeals found this supportive of the proposition that "`me-too' drugs carry the same qualities as generic drugs, irrespective of individual manufacturing differences, and therefore should be treated the same for purposes of the Drug Act." 585 F.2d at 584.

Mykocert was a condemnation case in which the plaintiff argued that its drugs should not have been seized, because they were not "new." Although the Mykocert court ultimately rejected plaintiff's claim, it stated:

Claimant's contention that Mykocert is not a new drug can only succeed if it is recognized by experts in the field as safe and effective. There are two possible alternatives that would substantiate claimant's assertions. Either a Mykocert type drug in its exact form, dosage, and application must be recognized by experts (even though the drug may be marketed under a different name) [a generic drug] or each of the component parts of Mykocert must be recognized with the critical caveat that the combination of these parts does not in any way create a new drug under [FDA regulations].

United States v. An Article of Drug . . . "Mykocert," supra, 345 F. Supp. at 575, quoted in United States v. Articles of Drug (Lannett), supra, 585 F.2d at 584 (emphasis and brackets added by Lannett court).

The Lannett court coupled this Mykocert definition with (1) the Supreme Court's indication in Hynson of similarities between generic products and the pioneer drugs they copy and (2) the Government's concession of identity between Lannett's products and their approved counterparts, and then concluded that summary judgment on the "new drug" issue should have been denied by the district court. 585 F.2d at 584.

As reflected in my findings of fact later set forth, the extensive testimony adduced by the parties in this proceeding has convinced me that adherence to the dictum of the court of appeals in Lannett would pose a substantial danger to public health, given the products here involved. Indeed, with all respect, I find the court's language at odds with the Act, and I share the views expressed by my colleague, Judge Meanor, in Pharmadyne Laboratories, Inc. v. Kennedy, 466 F. Supp. 100 (D.N.J.), aff'd on other grounds, 596 F.2d 568 (3d Cir. 1979), which involved facts matching those here, and which was decided shortly after Lannett. While the Pharmadyne opinion contains several powerful arguments in support of Judge Meanor's determination not to follow Lannett, of particular force is his discussion of the interrelationship between §§ 355 and 321(p) of the Act.

He first construed Lannett to mean that a generic copy of an approved drug product is not a "new drug" for purposes of section 321(p) "if its therapeutically active ingredients are identical to [the] recognized drug both chemically and quantitatively." Id. at 103. Thereafter, however, Judge Meanor did not confine his analysis of the meaning of "new drug" to section 321(p) alone. Stating that the Lannett court had "compartmentaliz[ed] the statute," id., Judge Meanor looked to the new-drug application section of the Act, section 355, "to shed light on the meaning of ["new drug"] as defined in § 321," id. Section 355(b) requires that all new-drug applications contain:

. . . (1) full reports of investigations which have been made to show whether or not such drug is safe for use and whether such drug is effective in use; (2) a full list of the articles used as components of such drug; (3) a full statement of the composition of such drug; (4) a full description of the methods used in, and the facilities and controls used for, the manufacture, processing, and packing of such drug. . . .

21 U.S.C. § 355(b) (1976) (emphasis added).

As indicated above, drug products generally contain much in addition to their active ingredients. See discussion of excipients (binders, fillers, disintegrants, etc.), supra, pp. 963-964. Judge Meanor stated that if these components were of concern to Congress for pioneer drugs, as they plainly were, see 21 U.S.C. § 355(b)(2) (1976), then "it is entirely unclear . . . why they would not be of similar concern in the manufacture of a `me-too' product." 466 F. Supp. at 104. Similarly, he suggested that if manufacturing methods, facilities, and controls are important considerations with respect to the pioneer, they were intended by Congress to be equally important with respect to the manufacture of that pioneer's generic counterparts. Id. When sections 321 and 355 are read together, the judge concluded, "it should be clear that Congress was concerned with such things as quality control, bioavailability and bioequivalence of the drugs prescribed for and consumed by the public." Id. ^6a

6a In addition, the court ascribed great weight to the FDA's interpretation of the Act, relying on the common rule of statutory construction that "where a statute is reasonably susceptible to alternative constructions, and one of those constructions has been given the statute by the agency charged with its administration and enforcement, the latter is to prevail." 466 F. Supp. at 104.
See Udall v. Tallman, 380 U.S. 1, 85 S.Ct. 792, 13 L.Ed.2d 616 (1965).
Finally, Judge Meanor quite properly distinguished Weinberger v. Hynson, Westcott Dunning, Inc., 412 U.S. 609, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973), relied on in Lannett, by pointing out that where a pioneer product's active ingredients are found unsafe or ineffective, then plainly generic versions of the pioneer containing the same active ingredients should be treated identically for withdrawal purposes; however, it is just as plain that the converse is not true. As indicated, a multitude of other factors beyond simple identity of active ingredients must be considered when making a determination, as an initial matter, that a given generic copy of an approved product is safe and effective. See 466 F. Supp. at 104 n. 7; discussion of bioavailability and bioequivalence, supra pp. 962-965; Findings of Fact 28-49.

Judge Meanor's analysis is persuasive, particularly for its reading of section 321(p) in conjunction with section 355. In addition, a reading of section 310.3 of FDA's regulations, upon which the Lannett court relied, also proves instructive.

As the Lannett court noted, FDA's regulations do not purport to give an exhaustive definition of "newness." The pertinent portion of section 310.3 reads:

(h) The newness of a drug may arise by reason ( among other reasons) of:

(1) The newness for drug use of any substance which composes such drug, in whole or in part, whether it be an active substance or a menstruum, excipient, carrier, coating, or other component.

(2) The newness for drug use of a combination of two or more substances, none of which is a new drug.

(3) The newness for drug use of the proportion of a substance in a combination, even though such combination containing such substance in other proportion is not a new drug.

(4) The newness of use of such drug in diagnosing, curing, mitigating, treating, or preventing a disease, or to affect a structure or function of the body, even though such drug is not a new drug when used in another disease or to affect another structure or function of the body.

(5) The newness of a dosage, or method or duration of administration or application, or other condition of use prescribed, recommended, or suggested in the labeling of such drug, even though such drug when used in other dosage, or other method or duration of administration or application, or different condition, is not a new drug.

21 C.F.R. § 310.3(h)(1)-(5) (1980) (emphasis added).

Such a qualified listing ("among other reasons") is hardly "powerful evidence" against FDA's position. In fact, I find strong support for the position of the United States in the regulation just quoted.

Section 310.3(h)(1) speaks of the "newness for drug use of any substance which composes such drug" (emphasis added). Excipients, carriers, and coatings are specifically mentioned. And "newness for drug use" contemplates a particular use, see 21 C.F.R. § 310.3(h)(5) (1980). This indicates that different inactive ingredients in a generic version than in its pioneer counterpart will constitute a "new use" if the generic version is used for the same medical indications as the pioneer, regardless of whether these inactive ingredients are generally recognized as safe and effective for other uses. This same argument holds true for the other sections of the regulation which speak of "newness for drug use." For example, § 310.3(h)(2) refers to "the newness for drug use of a combination of two or more substances, none of which is a new drug." Further, with regard to § 310.3(h)(2), that section speaks of a new "combination." This plainly indicates that if the active and inactive ingredients in Premo's products are not identical to those in the products they purport to copy, then Premo's products constitute a new "combination" under § 310.3(h)(2) and are considered "new drugs" under the regulations.

Additional support for the Government's position is contained in § 310.3(h)(3), which concerns "newness for drug use of the proportion of a substance in a combination" even though the same combination in other proportions would not be a "new drug." This is directly applicable to the general/new-drug issue since, even if a manufacturer produces a generic product that has ingredients identical to its pioneer counterpart, if the generic version is more, or less, bioavailable than its pioneer, it is effectively "a new proportion of a substance in combination" and thus considered a new drug under the regulations.

United States v. An Article of Drug . . . "Mykocert," supra, on which the Lannett court relied, need not be discussed in detail. The Mykocert court's construct for determining "old drug" status, quoted by the court in Lannett, is dictum only. The court found Mykocert a new drug under the Act, because there was convincing evidence that Mykocert was not generally recognized by experts as safe and effective. In fact, the language the court employed in announcing this determination is telling, and undercuts both the result in Lannett and defendants' position here. Presented with numerous affidavits by Government experts, all denying recognition of Mykocert's safety and effectiveness, the court stated:

[I]t cannot be denied that the affidavits of five of the leading doctors in the field which deny general recognition creates more than a "mere" conflict. It is inconceivable that a drug such as this could be considered generally recognized in the face of such learned non-recognition.

345 F. Supp. at 575 (emphasis added).⁷

7 In this regard, see discussion of testimony of Government's experts pp. 973-976 infra. See also Findings of Fact 10-21.

Shortly after the Pharmadyne decision, Premo Pharmaceutical Laboratories, Inc. v. United States, 475 F. Supp. 52 (S.D.N Y 1979), was decided. It establishes something of an intermediate position between Pharmadyne and Lannett. Premo involved an action for a declaratory judgment that plaintiff's Insulase — its generic version of the approved pioneer product, Diabinese, used in the treatment of adults with mild to moderate chronic diabetes — is not a "new drug" under section 321(p). The active ingredients in Diabinese and Insulase are the same; the inactive ingredients, however, are not. Id. at 53.

The district judge rejected Premo's argument that the word "drug" in the term "new drug" refers only to a drug product's active ingredient. 475 F. Supp. at 54-55. He likewise rejected the arguments proffered by FDA: (1) that any particular drug product is a "new drug" if it is not generally recognized as safe and effective; (2) that a drug product is a new drug if its excipients differ in any way from those in the approved product, even if the active ingredients are the same; and (3) that a drug product is a new drug if there is any difference whatever between the bioavailability curves of the two products. Id. at 55. Thus the court stated:

In interpreting the term "new drug", the Court is mindful that some differences in excipients do affect the safety or effectiveness of a drug product, but that others do not, and that Congress has entrusted the FDA, not the Court, with finally determining the safety and effectiveness of drugs.

Id. Yet, despite the court's assertion of congressional intent, it took extensive evidence and ultimately determined that Insulase is not a "new drug." In doing so, the court applied the following standard:

[W]hen the active ingredient in a questioned drug product is the same as the active ingredient in a drug product already on the market and generally recognized as safe and effective, and when the excipients in the two drug products are different, and when the excipients in the questioned product are generally recognized individually to be safe, the manufacturer of the questioned product is entitled to a declaration that its product is not a "new drug" within the meaning of 21 U.S.C. § 321(p), only if, the evidence has shown no reasonable possibility that differences between the excipients in the recognized and questioned products will make the questioned product less safe or effective than the recognized product.

Id. ⁸

8 The district court's decision in Premo was substantially followed by the United States District Court for the Southern District of Florida in United States v. Generix Drug Corp., 498 F. Supp. 288 (S.D.Fla. 1980). The standard enunciated by the Generix court does differ from that in Premo, however, in that it would find a drug product "new" within the meaning of section 321(p) "where the government demonstrates that there is some reasonable possibility that differences between the copied drug product and the generic drug product affect the safety and effectiveness of the generic drug product." Id. at 293. This appears to recognize the possibility that differences in manufacturing practices and/or sources of active ingredients can cause differences in bioavailability and thus in safety and effectiveness.

On appeal the Court of Appeals for the Second Circuit reversed. It held that the district court should not have considered the actual safety and effectiveness of Premo's generic product; this was a question exclusively for FDA resolution and the district court should have limited its inquiry to the narrow question whether there existed general expert recognition of the product's safety and effectiveness:⁹

Where a dispute exists as to whether a drug product is "generally recognized" by the experts to be safe and effective, the function of the district court is limited to determining that issue, not whether the product (including one claimed to be a "me-too" drug) is in fact safe and effective. The latter issue is to be determined by the FDA. . . .

Premo Pharmaceutical Laboratories, Inc. v. United States, supra, at 803.

9 A further question for judicial resolution, of course, is whether a particular product has been used "to a material extent or for a material time." See 21 U.S.C. § 321(p)(2) (1976); Premo Pharmaceutical Labs, Inc. v. United States, supra, at 800, 804. This question need be reached, however, only if there is no genuine dispute among experts as to a product's safety and efficacy, though the Premo court did note that "although Premo has produced and sold at wholesale some 16,500,000 Insulase tablets . . . there is no evidence that Insulase has been used to a material extent or for any substantial period of time." Id. at 804. The sales of Premo's products in question here range from 342,000 tubes of betamethasone valerate cream to 34.6 million tablets of hydroxyzine hydrochloride. See Plaintiff's Exhibit 64. None of Premo's products had been marketed for as much as two years at the time the Government brought this action; indeed, five of the eight had been marketed for less than a year. Id. There is nothing in the record that leads me to conclude that these figures constitute material use or use for a material time with respect to any of the products at issue. In fact, there is extensive evidence in the record that the relatively short time each of the products has been marketed is not sufficient to form any conclusion about the safety and effectiveness of these products. P. Ex. 1, ¶¶ 25-27 (Barr Affidavit); P. Ex. 3, ¶ 25 (Cortell Affidavit); P. Ex. 4, ¶ 30 (Crout Affidavit); P. Ex. 6, ¶ 19 (Perel Affidavit).

In addition to the "plain language of the statute," id. at 800, the Second Circuit based this conclusion on several grounds. The court found the history of the Act, referred to above, confirmed the "limited scope" of the general recognition exception to the "new drug" standard of § 321(p). Id. at 802. Moreover, the court derived support for its position from the legislative history of the 1961 Amendments to the Act. Citing the report of the Senate Committee on the Judiciary, quoted in the margin,¹⁰ the court stated: "[T]he legislative history of the 1961 Amendments suggests that Congress intended all new drug products to be subject to FDA premarket clearance procedures." Id. at 802 n. 7. "Nothing in the language of the Act or its legislative history," the court added, "suggests that it is the task of the courts to determine in the first instance whether a drug product is safe, effective or `therapeutically equivalent' to an already approved drug." Id. at 803.

10 The Senate Report stated of the new-drug provision that

[a]nything which currently required a new drug application — such as a new product or new use for an existing product which is not generally recognized as safe — will in the future have to make the test of effectiveness laid down in the bill as amended.

S.Rep. No. 1744, 87th Cong., 2d Sess. (1962), reprinted in [1962] U.S. Code Cong. Ad.News 2884, 2893 (emphasis added). The Second Circuit found its position further buttressed by the minority report:

The definition [of a "new drug"] encompasses every use of every truly new product never before marketed for any purpose because the product is clearly not generally recognized as safe and has not been used for a material time. . . . The law provides that any use of a brand new product must be cleared under the test of safety (plus effectiveness under the bill).

Id. at 804, quoted in Premo Pharmaceutical Labs, Inc. v. United States, supra, at 802 n. 7 (emphasis added).

Equally important to the court was FDA's scientific expertise:

[W]hether the product (including one claimed to be a "me-too" drug) is in fact safe and effective . . . is to be determined by the FDA which, as distinguished from a court, possesses superior expertise, usually of a complex scientific nature, for resolving the issue. The entire statutory scheme envisages that the FDA will perform the difficult task of investigation and scientific evaluation usually required to determine whether a drug product is safe and effective. . . .

. . . The rule that the FDA rather than the courts must first determine the safety and effectiveness of a drug is but an extension of the general principle that the agency is usually better equipped by reason of its expertise to make the determination than the court.

Id. at 803.

Finally, the Second Circuit rejected the district court's argument that adhering to the "general recognition" requirement of § 321(p) "would require FDA approval for all drug products" and "would frustrate the long-recognized purpose of the Act to allow the marketing of safe and effective `me-too' drug products without costly and time-consuming FDA approval," id. at 804 (quoting 475 F. Supp. at 55). The court noted that the "general recognition" exemption had permitted many drugs — generally recognized as safe prior to the Act's enactment in 1938 and as effective prior to the 1962 Amendments — to be marketed without filing NDAs. Id. at 805.

Thus, the court concluded, "the purpose of the Act is to subject all ["me-too"] drug products not generally recognized as safe and effective . . . to the premarket clearance requirements of the Act." Id. at 805. Because the excipients in the generic and pioneer products differed, however, the Second Circuit explicitly refrained from deciding "whether an identical, precise copy of an FDA-approved drug, which contains the exact same ingredients and excipients in the same proportions may successfully claim general recognition based on FDA-approval of the drug product which has been copied." Id. at 805 n. 9.

As to the underlying question of general expert recognition of the generic product at issue, the Second Circuit found dispositive (1) the district court's finding, made prior to the hearing conducted by that court on the Premo product's safety and effectiveness, that "[t]he specific combination of active drug and excipients which constitutes [the generic product] are not at this time generally recognized among experts . . . as safe and effective," id. at 801 (quoting district judge's denial of preliminary injunction), and (2) the inability of any of the experts who testified at trial, including Premo's witnesses, to testify that the generic product was generally recognized as safe and effective. Indeed, the studies made by Premo and offered as evidence of its product's safety and efficacy had not been published, id. at 800, and "were not available to the community of experts generally," id. at 804. With regard to this latter point — and significant for the present case — there is repeated language in the court of appeals' opinion that suggests the absence of published scientific material as to the generic product's safety and effectiveness would have been conclusive as to the general-recognition issue even absent testimony by the Government's experts. See id. at 800, 803, 805.¹¹

11 The court noted that "the studies offered as the basis for [Premo's experts'] opinion and essential to the determination of safety and effectiveness were made by Premo and not available to the community of experts generally." Premo Pharmaceutical Labs, Inc. v. United States, supra, at 804. And, as indicated, the Second Circuit concluded that the district court should have confined itself to determining whether " based on . . . published scientific studies or other publicly available data, scientific experts `generally recognized' Insulase to be safe and effective." Id. at 800 (emphasis added). Finally, the court flatly stated: "[T]he unawareness of the drug product by experts generally . . . preclude[s] its qualifying for exclusion as `generally recognized.'" Id. at 803 (emphasis added).

Moreover, even assuming the existence of such publicly available material, the court specifically found that "a genuine dispute among qualified experts regarding a drug product's safety and effectiveness preclude[s] its qualifying for exclusion as `generally recognized.'" Id. at 803. And on this issue there was a "sharp difference of opinion" among the experts who testified at trial. Id. at 804.

Thus, the Second Circuit held Premo's Insulase to be a "new drug" within the meaning of the Act, reversed the judgment of the district court, and remanded the case with directions to dismiss the complaint.

Premo stands not only as a rejection of the approach adopted by the district court in that case, and followed by the court in Generix, but is at odds with the dictum of the Court of Appeals for the Third Circuit in Lannett — if that case is read to require only general recognition of the safety and effectiveness of a drug product's active ingredients.

The Lannett court's dictum is, of course, entitled to considerable weight. United States v. Bell, 524 F.2d 202, 205-206 (2d Cir. 1975); Gabbs Exploration Co. v. Udall, 315 F.2d 37, 39 (1962).^11a However, the plain language of the Act and a close examination of the evidence presented in this case lead inexorably to the conclusion that the analysis advanced by the Second Circuit in Premo is correct: A drug product is a "new drug" under the Act unless it is "generally recognized, among experts . . . as safe and effective for use under conditions prescribed" and "used to a material extent or for a material time" under noninvestigative conditions. 21 U.S.C. § 321(p) (1976); Premo Pharmaceutical Laboratories, Inc. v. United States, supra, at 801.

11a Although this dictum must be accorded a measure of deference, it is not binding. "[J]udicial dictum by the Court uttering it . . . while of great significance and entitled to this Court's respect does not preclude . . . this Court from reaching its own decision after independent consideration and study of the question." United States v. Bell, 524 F.2d 202, 206 n. 4 (2d Cir. 1975) (quoting Perlman v. Timberlake, 172 F. Supp. 246, 253 (S.D.N.Y. 1959)). See Kastigar v. United States, 406 U.S. 441, 454-55 (1972); Powell v. McCormack, 395 U.S. 486, 519 n. 40, 89 S.Ct. 1944, 1962, 23 L.Ed.2d 491 (1969); Cohens v. Virginia, 19 U.S. (6 Wheat.) 264, 398 (1821); Boyd v. Henderson, 555 F.2d 56, 60 n. 5 (2d Cir. 1977).

Based on the extensive record developed before me, I am convinced that reliance on Lannett's dictum would be misplaced and that, had the court of appeals had before it in Lannett the record developed before me here, it would not have uttered such dictum. I also reject the approach of the district court in Premo. ¹²

12 In addition to the Lannett, Pharmadyne, Generix, and Premo decisions, a sixth case, United States v. AHC Pharmacal, Inc., No. 78-3685 (S.D.Fla. Mar. 2, 1979), is cited by defendants as involving the "new drug" issue in the generic/new-drug context. AHC involved drug products that were found by the court to be "new drugs," because they were misbranded and had not been manufactured in conformance with good manufacturing practices. In making its ruling, the court stated that if the products were relabeled and manufactured in conformance with good manufacturing practices, they would no longer be "new drugs" for the purposes of section 321(p). The opinion is unclear, however, whether the products at issue were generic copies of approved products or were, in fact, the approved products themselves and had simply been relabeled by the manufacturer. In any event, the language in the opinion suggesting that the products would no longer be "new drugs" if relabeled and manufactured in conformance with good manufacturing practices is pure dictum and adds nothing to the array of opinion discussed in text.

IV

Applying the Second Circuit's standard to the products at issue here, it is clear that the eight Premo products are "new drugs" for purposes of § 321(p); in fact, I need not even reach the question whether the products have been used to a material extent or for a material time,¹³ because I find that Premo's products are not generally recognized as safe and effective by qualified experts.

Considering first those products that the parties agree have inert ingredients not identical to those of their pioneer counterparts,¹⁴ the Government presented voluminous testimony from its expert witnesses ¹⁵ that the Premo products are not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof.

14 There is agreement on this issue as to all of the products but two, information pertaining to which is the subject of a protective order entered by this court. The two products are discussed, in general terms, in text infra.

15 The parties do not seriously dispute that each of the Government's expert witnesses is "qualified by scientific training and experience to evaluate the safety and effectiveness of drugs," 21 U.S.C. § 321(p) (1976). It is thus unnecessary to recite the credentials of these witnesses. The curriculum vitae of each was admitted into evidence, see P. Ex. 1 (Dr. Barr); P. Ex. 2 (Dr. Cabana); P. Ex. 3 (Dr. Cortell); P. Ex. 4 (Dr. Crout); P. Ex. 6 (Dr. Perel); P. Ex. 7 (Dr. Shangraw), and each was examined and found qualified to render expert testimony with regard to the products about which he testified. See Fed.R.Evid. 702.
Similarly, the defendants presented numerous expert witnesses ( see Defendants' Proposed Finding of Fact 6) whose qualifications and credentials, with the exception of Mr. Robillard, were not challenged by the United States.

As to Premo's Betamethasone valerate (Beta Val) cream, see P. Ex. 4, ¶ 23 (Crout Affidavit); P. Ex. 1, ¶ 24 (Barr Affidavit); P. Ex. 7, ¶ 16 (Shangraw Affidavit).

As to Premo's chlorthalidone tablets, see P. Ex. 4, ¶ 23 (Crout Affidavit); P. Ex. 1, ¶ 24 (Barr Affidavit); P. Ex. 3, ¶ 24 (Cortell Affidavit); P. Ex. 2, ¶¶ 41, 45 (Cabana Affidavit); P. Ex. 7, ¶ 16 (Shangraw Affidavit).

As to Premo's doxylamine succinate with pyridoxine HCL (Doxine) tablets, see P. Ex. 4, ¶ 23 (Crout Affidavit); P. Ex. 1, ¶ 24 (Barr Affidavit); P. Ex. 6, ¶ 18 (Perel Affidavit); P. Ex. 2, ¶¶ 44, 45 (Cabana Affidavit); P. Ex. 7, ¶ 16 (Shangraw Affidavit).

As to Premo's Hydroxyzine pamoate (Hy Pam) capsules, see P. Ex. 4, ¶ 23 (Crout Affidavit); P. Ex. 1, ¶ 24 (Barr Affidavit); P. Ex. 6, ¶ 18 (Perel Affidavit); P. Ex. 7, ¶ 16 (Shangraw Affidavit).

As to Premo's hydroxyzine hydrochloride tablets, see P. Ex. 4, ¶ 23 (Crout Affidavit); P. Ex. 1, ¶ 24 (Barr Affidavit); P. Ex. 6, ¶ 18 (Perel Affidavit); P. Ex. 7, ¶ 16 (Shangraw Affidavit).

As to Premo's trifluoperazine hydrochloride tablets, see P. Ex. 4, ¶ 23 (Crout Affidavit); P. Ex. 1, ¶ 24 (Barr Affidavit); P. Ex. 6, ¶ 18 (Perel Affidavit); P. Ex. 2, ¶¶ 42, 45 (Cabana Affidavit); P. Ex. 7, ¶ 16 (Shangraw Affidavit).¹⁶

16 There was no testimony adduced from any of defendants' experts on the question of general expert recognition as to these products themselves. Rather, defendants' experts directed their testimony to asserting that the active ingredient (or active ingredients in combination) of each of Premo's products is generally recognized as safe and effective among qualified experts, that the pioneer counterpart of each product is also so recognized, and that, for the most part, the active ingredients do not have known, medically significant bioequivalence problems. None of this testimony, of course, is relevant to the question now before me: whether each Premo product is generally recognized by qualified experts to be safe and effective.

While I do not believe the Act requires unanimity for a finding of general recognition, it is patent that there is here, at minimum, such "a genuine difference of medical opinion among . . . experts" on the question as to preclude a finding of general recognition of safety and effectiveness by experts with respect to the products listed above. See United States v. An Article of Drug — Furestrol, 294 F. Supp. 1307, 1311 (N.D.Ga. 1968), aff'd on the opinion below, 415 F.2d 390, 392 (5th Cir. 1969); Merritt Corp. v. Folsom, 165 F. Supp. 418, 421 (D.D.C. 1958). Moreover, all of the Government experts stated that they themselves do not recognize Premo's products as safe and effective. See P. Ex. 4, ¶ 23 (Court Affidavit); P. Ex. 1, ¶ 23 (Barr Affidavit); P. Ex. 3, ¶ 23 (Cortell Affidavit); P. Ex. 6, ¶ 17 (Perel Affidavit); P. Ex. 2, ¶¶ 35, 41-45 (Cabana Affidavit); P. Ex. 7, ¶ 15 (Shangraw Affidavit.)¹⁷ As the court put it in United States v. An Article of Drug . . . "Mykocert," supra, 345 F. Supp. at 575 (emphasis added);

17 These opinions were based in great part on the various factors that can affect the bioequivalence of two drug products that contain identical quantities of the same active ingredients. See discussion supra. See also Finding of Fact 21.

While we do not express an opinion as to whether a "mere" conflict in expert opinion constitutes lack of general recognition, it cannot be denied that the affidavits of five of the leading doctors in the field which deny general recognition creates more than a mere conflict. It is inconceivable that a drug such as this could be considered generally recognized in the face of such learned non-recognition.

See Premo Pharmaceutical Laboratories v. United States, supra, at 803 ("[A] genuine dispute among qualified experts regarding a drug product's safety and effectiveness preclude[s] its qualifying for exclusion as "generally recognized").

As to the remaining two Premo products, its allopurinol and triamterene with hydrochlorothiazide medications, the Government concedes that at least one is virtually identical to its pioneer counterpart. Government's Response to Defendants' Proposed Finding of Fact No. 99 (sealed). Defendants contend that the final Premo product is also identical to its pioneer, but this is disputed by the Government. Compare Defendants' Proposed Findings of Fact in Opposition to Plaintiff's Motion for Preliminary Injunction No. 138 (sealed) with Government's Responses to Defendants' Proposed Finding of Fact No. 138 (sealed). See also Government's Proposed Findings of Fact and Conclusions of Law 152-56, 204-06 [hereinafter Government's Proposed Findings] and defendants' critiques.

However, even assuming an identity of ingredients, qualitatively and quantitatively, between these two Premo products and their pioneers, the Government argues that (1) there is a potentially significant difference between the manufacturing processes¹⁸ used for one of the Premo products and its pioneer and between the two products' particle size,¹⁹ Government's Proposed Findings 153-54 (sealed); and (2) with respect to the second Premo product there is no information available as to the particle size and source²⁰ of Premo's active ingredient. Government's Proposed Finding 205 (sealed). The significant point here, of course, is not these differences themselves, both real and potential, but that, because of them, the Government witnesses have stated that there is no general expert recognition of the safety and efficacy of these two Premo products. As to allopurinol, see P. Ex. 4, ¶ 23 (Crout Affidavit); P. Ex. 1, ¶ 24 (Barr Affidavit); P. Ex. 7, ¶ 16 (Shangraw Affidavit). As to Premo's Triamthiazide, see P. Ex. 4, ¶ 23 (Crout Affidavit); P. Ex. 1, ¶ 24 (Barr Affidavit); P. Ex. 3, ¶ 24 (Cortell Affidavit); P. Ex. 2, ¶¶ 35, 45 (Cabana Affidavit); P. Ex. 7, ¶ 16 (Shangraw Affidavit).²¹ Nor do the Government experts themselves recognize Premo's allopurinol and triamterene with hydrochlorothiazide products as safe and effective. See P. Ex. 4, ¶ 23 (Crout Affidavit); P. Ex. 1, ¶ 23 (Barr Affidavit); P. Ex. 3, ¶ 23 (Cortell Affidavit); P. Ex. 2, ¶¶ 35, 43 (Cabana Affidavit); P. Ex. 7, ¶ 15 (Shangraw Affidavit).

18 As to the significance of differences in manufacturing practices, see Findings of Fact 31-49. Indeed, defendants concede that "[d]ifferences in manufacturing techniques can affect the performance of a drug." Defendants' Memorandum on Premo Pharmaceutical Laboratories, Inc. v. United States at 8.

19 As to the significance of differing particle sizes between two otherwise identical active ingredients, see Findings of Fact 29-30.

20 As to the significance of differing sources of the "same" active ingredient, see Findings of Fact 28-30; see generally id. 26-51.

21 One of defendants' experts, Dr. Tessler, testified that Premo's triamterene with hydrochlorothiazide product is generally recognized by experts as safe and effective. P. Ex. 86-B, at 50 (Tessler Deposition). This opinion, however, was based on Dr. Tessler's use of, and reading of, literature concerning the pioneer product Diazide, id. at 141-42, and the fact that the product has been allowed to be marketed by FDA, id. at 52.
There was no expert testimony from defendants' witnesses as to general expert recognition of the safety and effectiveness of Premo's allopurinol product. See note 16 supra.

As indicated, in Premo the Court of Appeals for the Second Circuit explicitly left unresolved the question whether "an identical, precise copy of an FDA-approved drug, which contains the exact same ingredients and excipients in the same proportions may successfully claim general recognition based on FDA-approval of the drug product which has been copied." Premo Pharmaceutical Laboratories v. United States, supra, at 805 n. 9. Nevertheless, I find the underlying logic of the Premo decision helpful on the issue. As defendants concede, differences in particle size and source of active ingredients can affect the safety and effectiveness of a drug product. See Defendants' Critique on Government's Proposed Findings of Fact No. 28-30. As defendants also concede, "[d]ifferences in manufacturing techniques can affect the performance of a drug." Defendants' Memorandum on Premo Pharmaceutical Laboratories, Inc. v. United States at 8. See Findings of Fact 31-49. Thus, there is a substantial question as to the safety and efficacy of Premo's products, which is one for FDA — unless there is general expert recognition of the safety and effectiveness of the products and they have been used to a material extent or for a material time. Here, again, however, there is no such general recognition; the testimony of the Government witnesses demonstrates, at the least, a "genuine difference of medical opinion among experts" and thereby precludes such a finding.²²

22 On this basis, too, I must reject even the narrow reading of the Lannett court's dictum, which would restrict the court's language solely to generic products identical both chemically and quantitatively to their pioneer counterparts. See note 4 supra.

V

Defendants raise a number of other arguments in support of their position. Of these, only a few merit even brief attention at this juncture.

Defendants contend that "FDA for over 30 years construed the statute so as to exclude generic versions of established drugs from the new drug definition." Defendants' Initial Post-Hearing Memorandum in Opposition to Plaintiff's Motion for Preliminary Injunction at 10. "Such a contemporaneous construction of a statute by an agency," defendants argue, "is clearly indicative of Congressional intent." Id. at 10-11. This contention is rejected. Regardless of what the agency's position has been in years past — and there has indeed been "a great deal of past waffling" on this point, Pharmadyne Laboratories, Inc. v. Kennedy, supra, 466 F. Supp. at 103 — it holds to that ground no longer. Indeed, I consider the Senate Report alluded to earlier, see note 11 supra, and the history of the Act itself, see p. 962 supra, as strong evidence that FDA's present interpretation accurately reflects the intent of Congress in enacting the Act.

Defendants also argue that the reading of section 321(p) adopted here renders superfluous the Act's "grandfather clauses, see, e.g., 21 U.S.C. § 321(p)(1)(1976)," thereby violating a venerable tenet of statutory construction. See Jackson v. Kelly, 556 F.2d 735, 740 (10th Cir. 1977); International Telephone Telegraph Corp. v. American Telephone Telegraph Co., 444 F. Supp. 1148, 1155 (S.D.N.Y. 1978) (every congressional enactment is purposive, and any interpretation that completely emasculates the words in a statute is improper). This argument, too, is without merit. It is answered by the Second Circuit in Premo, rejecting the district court's contention that the "general recognition" standard "would require FDA approval for all drug products." See Premo Pharmaceuticals, Inc. v. United States, supra, at 804.

Finally, defendants suggest the options available to FDA absent preclearance adequately protect the public. They do not. FDA's Good Manufacturing Practice (GMP) regulations, 21 C.F.R. § 210.1 et seq. (1980), are not sufficient to avoid the need for preclearance. The GMP regulations are intended to provide general, industrywide standards for the manufacture of drugs, not monitoring of individual products. P. Ex. 4 ¶ 31 (Crout Affidavit). Nor can FDA inspections provide an effective alternative to the preclearance process. Under the law there may be a time lag of up to six months between initial marketing of a drug and the time the drug is listed with FDA. See 21 U.S.C. § 360(j)(2)(1976). Thus, absent preclearance FDA may not even be aware that a product is on the market until it is too late to protect the public. P. Ex. 4 ¶ 32 (Crout Affidavit).

Moreover, the adulteration provisions of 21 U.S.C. § 351 (1976), which FDA can invoke against imperfectly manufactured drugs, also fall short of ensuring the margin of safety provided by preclearance. Most significant in this regard is the point made by the district court in Pharmadyne. If a specific product is not itself approved — but rather marketed as an "old drug" on the basis of FDA-approval of its pioneer — and is suddenly found to have disastrous and unanticipated side effects, FDA would have no NDA or ANDA to revoke and would have to resort to the infinitely slower alternative of individual condemnation actions. I share Judge Meanor's doubt "that in such an emergency Congress intended that some drugs could be withdrawn by immediate disapproval of an NDA while their "me-too" counterparts could only be forced from the market by condemnation." Pharmadyne Laboratories, Inc. v. Kennedy, supra, 466 F. Supp. at 106.

For the same reasons, I am not persuaded by arguments that this is a matter best left to Congress. Congress has already acted, in 1938 and again in 1962, and has mandated that a tragedy such as that involving the "Elixir of Sulfanilamide" does not recur.

VI

Premo's eight products are "new drugs" for purposes of 21 U.S.C. § 321(p) (1976). As such, marketing of these products absent premarket clearance by FDA is in clear violation of the Federal Food, Drug, and Cosmetic Act.²³

23 Shipping new drugs without approved NDAs violates 21 U.S.C. § 355(a) (1976), which states:

No person shall introduce or deliver for introduction into interstate commerce any new drug, unless an approval of an [NDA] is effective with respect to such drug.

Violation of 21 U.S.C. § 355 (1976) is a prohibited act, id. § 331(d). Manufacturing, packing, and labeling unapproved "new drugs" violates section 331(k), which prohibits the doing of any act with respect to a drug if that act is done after the article (or its components) have traveled in interstate commerce and that act results in the articles being misbranded. A drug is misbranded if it is a prescription drug that is an unapproved new drug, because a prescription drug cannot bear the adequate directions for use required by statute, id. § 352(f)(1), and the lack of an approved NDA means that there is no FDA exemption from the adequate directions for use requirement. See generally United States v. An Article of Drug . . . "Mykocert," 345 F. Supp. 571, 573 (N.D.Ill. 1972).

There remains only the question of relief. Although defendants have advanced arguments to the contrary, it is clear to me that this court has the power to enjoin present and future violations of the Act, 21 U.S.C. § 332(a) (1976), solely on the basis that such violations have been established. United States v. Diapulse Corp. of America, 457 F.2d 25, 27-28 (2d Cir. 1972); United States v. Nutrition Service, Inc., 347 F.2d 233 (3d Cir. 1965), aff'ing on the basis of the opinion below, 227 F. Supp. 375, 388-89 (W.D.Pa. 1964); FTC v. Rhodes Pharmacal Co., 191 F.2d 744, 747 (7th Cir. 1951), rev'd in part on other grounds, 348 U.S. 940 (1955); Shadid v. Fleming, 160 F.2d 752, 753 (10th Cir. 1947); Henderson v. Burd, 133 F.2d 515, 517 (2d Cir. 1943); United States v. Generix Drug Corp., supra, at 293; United States v. Articles of Food Drug, 441 F. Supp. 772, 775 (E.D.Wisc. 1977); United States v. Medwick Laboratories, Inc., 416 F. Supp. 832, 833 (N.D.Ill. 1976).

Therefore, subject to an alternative form of order agreed upon by the parties and approved by this court, defendants will be enjoined, as the Government requests, from (1) shipping new drugs in interstate commerce and (2) manufacturing, processing, packing, and labeling any new drugs after shipment of their components in interstate commerce, unless the drugs are approved by FDA or are used in investigations in compliance with applicable regulations.

As to the questions of recall, destruction of in-process and finished stocks, and costs and inspection if recall and destruction are ordered, I desire additional argument and will set this matter down at an early date for such argument.

Accordingly, a form of order in accordance with this opinion should be submitted within ten days.

FINDINGS OF FACT Preliminary Statement

Following the hearing, at the direction of the court, the parties submitted and exchanged proposed Findings of Fact and Conclusions of Law and thereafter, again at the direction of the court, submitted comments upon each other's proposed findings. Subsequently, each party submitted a reply to these comments. All proposed findings of fact, comments, and replies were annotated to the record, and read by me. Where there was disagreement, my resolution of such disagreement was facilitated by the record references supplied by counsel. In my Findings I have retained those record references to aid appellate review.

In many instances, the parties were in agreement with each other's submissions. In other instances, there was agreement in terms of a proposed finding but disagreement as to its relevancy. The former has been indicated by me with the letter (A); the latter is indicated by (NDBI), standing for "not disputed but irrelevant."

This court is mindful of the fact that findings of fact under Federal Rule of Civil Procedure 52 should be the findings of the court and not the findings of counsel. While I have drawn heavily upon the proposed findings of the parties, and in many instances have adopted the language used, the findings are the court's, not counsel's. For example, as I have indicated above, I have carefully reviewed all of the proposed findings submitted by the United States, read those findings in the light of the critique submitted by the defendants, also giving consideration to the reply of the United States to such critique, and have, in each instance where there is a dispute, carefully reviewed the record references to resolve that dispute.

Additionally, I have reviewed the proposed findings of fact submitted by the defendants, the United States' critique thereof, and the defendants' reply to the government's critique. Again, where I found a dispute between the parties, I used the record references to resolve it.

For a better understanding of these findings, I have adopted the headings and subheadings of the proposed findings of the United States.

I. Background

1. Defendant Premo Pharmaceutical Laboratories, Inc., ("Premo") is a corporation which trades and does business at 111 Leuning Street, South Hackensack, New Jersey 07606, within the jurisdiction of this Court (Complaint, ¶ 2; Amended Answer, 5th Defense, ¶ 2). (A)

2. Defendant Seymour N. Blackman is president of Premo Pharmaceutical Laboratories and as such, has over-all responsibility for and authority over the operation of the firm, including, but not limited to, the methods, manner, and conditions of manufacture of the firm's products, and whether to submit new-drug applications and whether to ship products for which such applications have not been submitted or approved. He performs his duties at 111 Leuning Street, South Hackensack, New Jersey 07606, within the jurisdiction of this court. (Complaint, ¶ 3; Amended Answer, 5th Defense, ¶ 3). (A)

3. Defendants have been and are now engaged at their plant in 111 Leuning Street, South Hackensack, New Jersey, in manufacturing, processing, packing, labeling, and distributing in interstate commerce various articles of drug. Such drugs include, but are not limited to: allopurinol tablets, betamethasone valerate cream, chlorthalidone tablets, doxylamine succinate with pyridoxine HCl tablets, hydroxyzine pamoate capsules, hydroxyzine hydrochloride tablets, and trifluoperazine tablets. (Complaint, ¶ 4; Amended Answer, 5th Defense, ¶ 4). (A) Premo also manufactures some such products through a related company, Federal Pharmacal, Inc., in St. Croix, the Virgin Islands (P. Ex. 25-A, page 6; P. Ex. 25-B at 69-70 (Deposition of J. Blackman); Tr. at 2111 (Silverang); P. Ex. 72-B at 15-16 (Deposition of N. Martin)).

4. Defendants manufacture, process, pack, and label these drugs while they are held for sale after shipment of one or more of their components in interstate commerce (P. Ex. 64, 9, 10, and 11). (A)

5. Defendants have manufactured, processed, packed, labeled, and distributed in interstate commerce capsules containing triamterene with hydrochlorothiazide (Complaint, ¶ 4; Amended Answer, 5th Defense, ¶ 4), and will do so again if not enjoined (P. Ex. 85-A, page 2; P. Ex. 85-B at 119 (Deposition of S. Silverang)). (A)

6. Defendants' acts of manufacturing, processing, packing and labeling capsules containing triamterene with hydrochlorothiazide were done while those capsules were held for sale after shipment of one or more of their components in interstate commerce (P. Ex. 64, 11). (A)

7. There is no new drug application, approved by the Food and Drug Administration (FDA) pursuant to 21 U.S.C. § 355(a), on file for any of the drugs listed in ¶¶ 3 and 5, supra (Complaint, ¶ 6; Amended Answer, 5th Defense, ¶ 6).

Premo holds IND's (Investigation New Drug exemptions) for allopurinol, 100 and 300 mg, betamethasone valerate, and trifluoperazine, 2 and 5 mg.

Premo sought investigational new-drug exemptions for these products after the complaint in this action was filed. With respect to two of these products (trifluoperazine and betamethasone valerate), Premo has been told by the FDA that it may not begin research until further information has been submitted. Premo has stipulated that it has distributed all of these products commercially (P. Ex. 64), which distribution would violate the Act even if use of the drugs were approved for investigational purposes.

8. Defendants will, in the absence of the entry of injunctive relief, manufacture and distribute the drugs in question.

9. Defendants have, in addition to those drugs set out in ¶¶ 3 and 5, also manufactured and shipped other unapproved new drugs (P. Ex. 5 (Second Aff. of Dr. Crout), ¶ 9).

II. Premo's Products Lack Expert General Recognition Under 21 U.S.C. § 321(p)

10. Premo's allopurinol tablet is not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof (P. Ex. 4 (Aff. of Dr. Crout), ¶ 23; P. Ex. 1 (Aff. of Dr. Barr), ¶ 24; P. Ex. 3 (Aff. of Dr. Cortell), ¶ 24; P. Ex. 2 (Aff. of Dr. Cabana), ¶ 43; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 16).

11. Premo's betamethasone valerate cream is not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof (P. Ex. 4 (Aff. of Dr. Crout), ¶ 23; P. Ex. 1 (Aff. of Dr. Barr), ¶ 24; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 16).

12. Premo's chlorthalidone tablet is not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof (P. Ex. 4 (Aff. of Dr. Crout), ¶ 23; P. Ex. 1 (Aff. of Dr. Barr), ¶ 24; P. Ex. 3 (Aff. of Dr. Cortell), ¶ 24; P. Ex. 2 (Aff. of Dr. Cabana), ¶¶ 41, 45; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 16).

13. Premo's doxylamine succinate with pyridoxine HCl (Doxine) tablet is not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof (P. Ex. 4 (Aff. of Dr. Crout), ¶ 23; P. Ex. 1 (Aff. of Dr. Barr), ¶ 24; P. Ex. 6 (Aff. of Dr. Perel), ¶ 18; P. Ex. 2 (Aff. of Dr. Cabana), ¶¶ 44, 45; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 16).

14. Premo's hydroxyzine pamoate capsule is such that such drug is not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof (P. Ex. 4 (Aff. of Dr. Crout), ¶ 23; P. Ex. 1 (Aff. of Dr. Barr), ¶ 24; P. Ex. 6 (Aff. of Dr. Perel), ¶ 18; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 16).

15. Premo's hydroxyzine hydrochlorothiazide tablet is not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof. (P. Ex. 4 (Aff. of Dr. Crout), ¶ 23; P. Ex. 1 (Aff. of Dr. Barr), ¶ 24; P. Ex. 6 (Aff. of Dr. Perel), ¶ 18; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 16).

16. Premo's trifluoperazine tablet is not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof (P. Ex. 4 (Aff. of Dr. Crout), ¶ 23; P. Ex. 1 (Aff. of Dr. Barr), ¶ 24; P. Ex. 6 (Aff. of Dr. Perel), ¶ 18; P. Ex. 2 (Aff. of Dr. Cabana), ¶¶ 42, 45; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 16).

17. Premo's triamterene with hydrochlorothiazide (Triamthiazide) capsule is such that such drug is not generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use under the conditions prescribed, recommended, or suggested in the labeling thereof (P. Ex. 4 (Aff. of Dr. Crout), ¶ 23; P. Ex. 1 (Aff. of Dr. Barr), ¶ 24; P. Ex. 3 (Aff. of Dr. Cortell), ¶ 24; P. Ex. 2 (Aff. of Dr. Cabana), ¶¶ 35, 45; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 16).

18. As agreed by Premo's expert, Dr. Rhodes, it is necessary to have more information with respect to a drug product than simply the identity of its active ingredients for that product to be generally recognized as safe and effective (Tr. at 1417). (A)

19. As Dr. Rhodes conceded, the excipients used in Premo's drug products are not known to experts generally (Tr. at 1417-19, 1486 (Dr. Rhodes)). The details of the manufacturing practices used in the manufacture of Premo's drugs are not known generally to experts either (Tr. at 1419 (Dr. Rhodes)). In addition, quality control information with respect to the procedures followed by Premo is not generally known to experts ( id.; Tr. at 1486) (Dr. Rhodes)). It is not generally known by scientists what the source of the inactive ingredients is for Premo products (Tr. at 1419-20 (Dr. Rhodes)).

20. Without reliable biological availability, clinical trials and quality control data, and without knowing the excipients of a product, a product cannot be recognized as safe and effective even though the active drug substance is pure (Tr. at 1497-98 (Dr. Rhodes)).

21. In addition to stating their expert testimony that there is a lack of general recognition among experts of the safety and effectiveness of the Premo products named as examples in the complaint, a number of well qualified experts in the evaluation of drug safety and effectiveness have stated that they themselves do not recognize those drugs as safe and effective (P. Ex. 4 (Aff. of Dr. Crout), ¶ 23; P. Ex. 1 (Aff. of Dr. Barr), ¶ 23; P. Ex. 3 (Aff. of Dr. Cortell), ¶ 23; P. Ex. 6 (Aff. of Dr. Perel), ¶ 17; P. Ex. 2 (Aff. of Dr. Cabana), ¶¶ 35, 41, 42, 43, 44, 45; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 15).

These contentions were not withdrawn at trial nor were they weakened or diluted by cross-examination of the affiants.

III. There Is a Clear Potential for Differences in Bioavailability Between Products Purporting To Contain the Same Active Ingredients

22. "Bioavailability" is a scientific term used to refer to "the rate and extent to which the active drug ingredient or therapeutic moiety is absorbed from a drug product and becomes available at the site of drug action," 21 C.F.R. § 321.1(a) (P. Ex. 4 (Aff. of Dr. Crout), ¶ 9; P. Ex. 1 (Aff. of Dr. Barr), ¶ 9). The term is also used to refer to the rate and extent to which an active drug ingredient is absorbed from a drug product and becomes available in the patient's blood stream (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 7; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 12; Tr. at 472 (Dr. Cabana); Tr. at 882 (Dr. Barr)), because it is assumed that equal blood levels of a drug will produce equal effects (Tr. at 26 (Dr. Crout)). (A)

23. The sciences of biopharmaceutics and pharmacokinetics, which are concerned with the bioavailability of drug products, are relatively young sciences. They were developed because of the need to evaluate different kinds of dosage forms, particularly as some of the older drugs were coming off patent so that additional companies were beginning to market them and, also, because analytical developments made possible the measurement of small amounts of drugs in the blood, thus permitting bioavailability studies (Tr. at 883 (Dr. Barr)). (A)

24. Bioavailability problems generally occur only when there is more than one product containing a particular active ingredient on the market (Tr. at 885 (Dr. Barr)).

25. Bioequivalence implies that the relative rate and extent of absorption of two drug products are sufficiently equal that one would expect the two drug products to behave, for all practical purposes, the same in a clinical situation (Tr. at 85 (Dr. Barr); see also Tr. at 473 (Dr. Cabana)). (A)

A. Inactive Ingredients

26. Two products, each containing the same active ingredients, can have a different bioavailability because of differences in their inactive ingredients (P. Ex. 1 (Aff. of Dr. Barr), ¶ 10; P. Ex. 2 (Aff. of Dr. Cabana), ¶¶ 23, 25, 30, 32; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13-B; P. Ex. 4 (Aff. of Dr. Crout, ¶¶ 9, 28, 29; Tr. at 472-73 (Dr. Cabana); Tr. at 1288, 1485 (Dr. Rhodes)) and thus can differ in safety and effectiveness (Tr. at 1414-15 (Dr. Rhodes)).

27. There are a variety of different inactive ingredients that may be present in a drug product, including diluents, binders, disintegrants, and surfactants. In addition, a drug may or may not be prepared with inactive ingredients in a solid solution. Each of these different ingredients and this potential difference in preparation of the active ingredients with inactive ingredients can be the cause of differences between two drug products, each containing the same active ingredients, which have different inactive ingredients (P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13-B). (A)

B. Active Ingredients

28. The "same" active ingredient from different sources may cause different effects in terms of bioavailability (P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13-A; P. Ex. 1 (Aff. of Dr. Barr), ¶ 10; P. Ex. 2 (Aff. of Dr. Cabana), ¶¶ 24, 32; P. Ex. 4 (Aff. of Dr. Crout), ¶¶ 9, 28; Tr. at 282-83 (Dr. Crout)). (A)

29. Active drug ingredients from different manufacturers may differ from each other in the following respects: polymorphism, that is, different crystal forms of the same ingredient that are chemically indistinguishable from each other but which may differ significantly with respect to properties such as density, melting point, solubility, the rate at which the active ingredient will go into solution, and stability (P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13A.1; Tr. at 214 (Dr. Crout); Tr. at 472 (Dr. Cabana)); the state of the active ingredient, that is, for some active ingredients the difference between a possible amorphous or non-structured state and a structured or crystal state (P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13-A.2); the association of the active ingredient with different solvents ( id., ¶ 13-A.3; Tr. at 716 (Dr. Shangraw)); the form of the dissolution properties of the product (P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13-A.4); particle size and distribution of the active ingredient, which may affect the rate and extent of dissolution of the active ingredient ( id., ¶ 13-A.5, Tr. at 214 (Dr. Crout); Tr. at 716 (Dr. Shangraw)). (NDBI)^1b

1b Unless indicated otherwise, "NDBI" will refer to the fact that a Finding is as it was proposed by the government and that the defendants' response to it had been "Not Disputed But Irrelevant."

30. When the particle size of drugs becomes smaller the surface area of the drug is increased, giving a greater opportunity for dissolution of the drug, because dissolution of the drug in the gastric media and intestinal media is a function of surface area (Tr. at 720 (Dr. Shangraw)). For example, grinding granular sugar into powdered sugar would make the sugar dissolve more quickly (Tr. at 712 (Dr. Shangraw)). Breaking particle size down to very minute quantities is termed "micronization" ( id.). (NDBI)

C. Manufacturing Variables

31. Even if the characteristics of the active ingredient are the same, and even if the same additives are used, two drug products manufactured by different manufacturers can vary in bioavailability because of differences in the manufacturing practices used (P. Ex. 1 (Aff. of Dr. Barr), ¶ 10; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13-C; P. Ex. 2 (Aff. of Dr. Cabana), ¶ 32; P. Ex. 4 (Aff. of Dr. Crout), ¶¶ 9, 28, 29; Tr. at 213-14, 282-83 (Dr. Crout); Tr. at 473 (Dr. Cabana); Tr. at 723-24 (Dr. Shangraw); Tr. at 1415 (Dr. Rhodes)). (NDBI)

1. Tablets

32. A tablet is made up of a number of ingredients, almost all of which can affect the bioavailability of the drug: the active ingredient, the filler, which is used to provide sufficient volume for the dosage form, disintegrating agents used in tablets to make them break apart, lubricants, used to ease the manufacture of the tablet, and binders (Tr. at 716-17 (Dr. Shangraw)). (NDBI)

33. Fillers can be soluble like lactose, or can be relatively insoluble like calcium sulfate, which is soluble in acid but not as soluble in alkaline or aqueous media. If the filler is not soluble, then it is more difficult for the drug to be released (Tr. at 719-20 (Dr. Shangraw)). Dissolution of a drug product is required for absorption (Tr. at 720 (Dr. Shangraw)). (NDBI)

34. In making a tablet it is very often necessary to "granulate," that is, to glue the small particles of the drug together with a binder to form a substance that will flow in the machine, or die, used to make the tablet and that will compress it (Tr. at 737-38 (Dr. Shangraw)). (NDBI)

35. A lubricant is often used in a tablet to help eject the tablet from the die after it has been made (Tr. at 738 (Dr. Shangraw)). (NDBI)

36. There are three basic manufacturing techniques for tablets: 1) wet granulation, in which a liquid glue is added to the powders to form a moist mass, which is then passed through a screen to break it down again; 2) dry granulation (also known as "slugging") in which the powders are made into a large crude tablet and then ground up into particles that will be larger than the powder so that they will flow in the machines; and, 3) direct compression, in which the drug ingredients are blended together to make the tablet directly without going through a granulation process (Tr. at 738-39; 723-24 (Dr. Shangraw)). In each case, the process is designed to produce a material that will flow in the machinery and that will be coherent when pressure is applied to it (Tr. at 739, 722 (Dr. Shangraw)). (NDBI)

37. Examples of the variables in manufacturing practices for tablets that may affect bioavailability also include method of disintegrant incorporation, choice of binder and granule hardness, granule size and size distribution, flow properties of the powder mixture, coating method, mixing time, presence of surfactant and its method of incorporation, compression pressure and compression dwell time, composition and concentration of disintegrants, lubricants and other excipients, the drug-to-excipient ratio, and the order of mixing drug excipients (P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13-C.1). (NDBI)

2. Capsules

38. For capsules, manufacturing variables that may affect bioavailability include methods used to reduce bulk for encapsulation, granule/powder size and size distribution, disintegrant/dispersents and method of incorporation of disintegrants or dispersents, lubricants and methods of incorporation of lubricants, surfactants and methods of incorporation of surfactants, the compression applied during filling of the capsules, flow properties of the powder mixture, and mixing time (P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13-C.2). (NDBI)

39. The hard shell of a capsule is made up of almost 90% gelatin but may also have small amounts of a plasticizer like propylene glycol or glycerine together with about 10% water and occasionally a little bit of sucrose to form a gelatin film that is fairly hard and brittle (Tr. at 743 (Dr. Shangraw)). (NDBI)

40. The contents of the gelatin capsules include the drug itself and fillers such as those used in tablets, lubricants to improve the flow of the material in the machinery and to keep the machinery from binding during manufacture. In some of the newer automatic machines the capsule contents are first made by forming a plug of powder in a cylinder and then ejecting that plug of powder into the capsule. The process is similar to making a tablet except that the plug in the capsule does not have nearly the density of a tablet (Tr. at 744 (Dr. Shangraw)). (NDBI)

41. Because disintegrating agents do not work in capsules as they do in tablets, it is often necessary to use a wetting agent or some surface active agent instead to help disperse the powder mass in the capsule (Tr. at 759). (NDBI)

42. The two factors that most influence the drug dissolution from a capsule and thus its bioavailability are 1) the properties of the drug, i.e., how soluble it is and how hydrophilic the surfaces of the drug particles are, meaning whether they wet well when they are exposed to water, and 2) the lubricant added to the drug material and how that lubricant is distributed in the powder mass (Tr. at 745; see also Tr. at 725 (Dr. Shangraw)). (NDBI)

43. The lubricant used in a capsule generally will coat the powder, i.e., the drug particles, and particularly with magnesium stearate, given enough blending time, will actually cover every surface of every particle of the drug (Tr. at 745, 769-70 (Dr. Shangraw)). In fact, Dr. Lerk has photographed a particle of sodium chloride which, after eventual dissolution, has left an unfolded sheath of magnesium stearate which had previously enclosed it (P. Ex. 28, page 324, Figure 10).

44. Coating the particles of drug with a lubricant will slow down dissolution of the drug particles (Tr. at 746, 764 (Dr. Shangraw); P. Ex. 30 at 1215). Magnesium stearate's effect on the dissolution of a drug is increased by increased blending time as shown by the work of Dr. Lerk (Tr. at 769-71 (Dr. Shangraw); P. Ex. 28 at 317 and 324, Figure 9). (NDBI)

45. Increasing the time of blending of the drug during its manufacture will more uniformly distribute the lubricant and will therefore slow down the dissolution of the drug (Tr. at 746, 764-65 (Dr. Shangraw); P. Ex. 30 at 1216, Figures 1 and 2). (NDBI)

46. When, in manufacturing a capsule, one increases the concentration of lubricant, the rate of dissolution of drug from that capsule can be dramatically decreased (Tr. at 754, 765 (Dr. Shangraw); P. Ex. 29, page 170, Figure 1, 171; P. Ex. 30 at 1217, Figure 3; cf. Tr. at 1297 (Dr. Rhodes)). A variation in formulation as small as a difference between 2% rather than 1% of magnesium stearate can cause significant differences in bioavailability as Dr. Barr's research has shown, differences great enough to cause a product Dr. Barr studied to be subtherapeutic in some patients (Tr. at 1010-11 (Dr. Barr)). (NDBI)

47. Both the type of filler that is used in a drug and the fluid in which the drug is dissolved can affect dissolution. Thus a filler like dicalcium phosphate, since it is not highly soluble in water but is soluble in acid, when mixed with a drug, will lead to fairly fast dissolution in an acid-like medium such as that found in the stomach and relatively slow dissolution in plain water or in an environment such as that found in the lower GI tract (Tr. at 755-56 (Dr. Shangraw); P. Ex. 29, page 172, Figure 7). (NDBI)

48. Magnesium stearate is a lubricant which forms a film on the surfaces of metals and prevents drug powders from sticking to those surfaces in the manufacture of a drug. Because magnesium stearate is "hydrophobic", i.e., it does not wet with water well, it slows down the dissolution of a drug in water (Tr. at 757 (Dr. Shangraw); P. Ex. 29, page 172 Figure 7). (NDBI)

49. How hard one packs the powder mass into a capsule can influence the dissolution of the drug out of that mass (Tr. at 758 (Dr. Shangraw); P. Ex. 29, Figure 10). (NDBI)

3. Topical Preparations

50. While FDA waives the requirement for submission of in vivo bioavailability data for topical preparations, variables that may affect bioavailability include the nature of the vehicle, particle size of the drug, solubility of the drug in the vehicle, and the presence or absence of surfactants (P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13-C.3).

D. Environmental Factors During Manufacture and Storage

51. Environmental factors can also cause differences between two otherwise identical drug products manufactured by different manufacturers (P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 13-D).

The government's point is simply that environmental factors can differ from manufacturing facility to manufacturing facility. A drug manufacturer must identify the manufacturing facility that it will use in any new drug application or abbreviated new drug application it files, see 21 C.F.R. § 314.1(c)(2), Form FD-356H(8)(a). A drug manufacturer wishing to change its manufacturing facility for a drug must file a supplemental application, 21 C.F.R. § 314.8(a)(4)(iv). The FDA requires bioavailability studies showing bioequivalence of the product manufactured in the new facility to the product manufactured in the approved facility prior to approval of such a supplemental application in appropriate circumstances, 21 C.F.R. § 320.21(b)(1). These requirements result in protection of the public from the differences that may occur in different facilities of the same manufacturer when that manufacturer is complying with the new drug application requirement but is not protected from the differences that may be caused by environmental variations between the facilities of the approved manufacturer and those of a manufacturer who markets without approval.

IV. Continued Marketing by Premo of Products Not Shown to be Bioequivalent to the Products They Copy Potentially Endangers the Public

52. In the normal course of business Premo's products can be expected to be substituted by pharmacists for the approved drug products that have the same active ingredients (P. Ex. 4 (Aff. of Dr. Crout), ¶ 14; P. Ex. 1 (Aff. of Dr. Barr), ¶ 18; P. Ex. 3 (Aff. of Dr. Cortell), ¶ 17; P. Ex. 2 (Aff. of Dr. Cabana), ¶ 39-C(1)). (A)

53. Physicians initially prescribing some of the approved drug products which Premo copies may be expected to carefully adjust the dosage for each patient (P. Ex. 4 (Aff. of Dr. Crout), ¶ 15; P. Ex. 6 (Aff. of Dr. Perel), ¶ 18). They may do so by monitoring a pharmacological response, for example, by checking uric acid levels for allopurinol ( see, e.g., Tr. at 393 (Dr. Bloom)) and potassium levels for triamterene plus hydrochlorothiazide ( see, e.g., Tr. at 98 (Dr. Crout); cf. Tr. at 430 (Dr. Bloom)).

54. If the Premo product is more bioavailable than the drug for which it is substituted, the resulting effective overdosage of the drug may, with respect to some products, present a risk for the patient (P. Ex. 4 (Aff. of Dr. Crout), ¶¶ 10, 15-17; P. Ex. 3 (Aff. of Dr. Cortell), ¶¶ 18-19; P. Ex. 6 (Aff. of Dr. Perel), ¶¶ 9-10; P. Ex. 2 (Aff. of Dr. Cabana), ¶ 39-C, D).

55. If a difference in bioavailability between Premo's drug and the approved product containing the same ingredients results in an effective underdosage of the active ingredient of the drug product in patients, that result could lead to therapeutic failures that, for some products in some patients, could constitute a risk for patients (P. Ex. 4 (Aff. of Dr. Crout), ¶¶ 10-14; P. Ex. 3 (Aff. of Dr. Cortell), ¶¶ 20, 22; P. Ex. 6 (Aff. of Dr. Perel), ¶¶ 11, 16; P. Ex. 2 (Aff. of Dr. Cabana), ¶ 39).

56. Unless a physician is alerted to look for a drug-related therapeutic failure from low bioavailability or adverse reactions from high bioavailability, he may mistakenly attribute the problem he sees to the disease state being treated (P. Ex. 3 (Aff. of Dr. Cortell), ¶ 25; P. Ex. 1 (Aff. of Dr. Barr), ¶¶ 25-27; P. Ex. 6 (Aff. of Dr. Perel), ¶ 19; P. Ex. 4 (Aff. of Dr. Crout), ¶ 30; Tr. at 899, 905-07 (Dr. Barr)). For this reason, the relatively short amount of time each of the Premo products has been marketed ( see P. Ex. 64) makes it impossible to draw a valid conclusion about the safety and effectiveness of those drugs (P. Ex. 3 (Aff. of Dr. Cortell), ¶ 25; P. Ex. 6 (Aff. of Dr. Perel), ¶ 19; P. Ex. 1 (Aff. of Dr. Barr), ¶¶ 25-27; P. Ex. 4 (Aff. of Dr. Crout), ¶ 30).^2a

2a I reject, in this respect, as unpersuasive, the testimony of Dr. Bloom (T-417-20) as tendered by the defendants.

57. It is possible for a drug to exhibit both a shallow dose-response curve and a steep dose-response curve at the same time: Although with some drugs an increase in dosage will not result in increased therapeutic effect (yielding a shallow curve), the same increase in dosage may result in an increased toxic effect (yielding a steep curve) (Tr. at 928-33 (Dr. Barr); P. Ex. 63). In other words, there could be an increase in toxic or adverse effects in response to an increase in effective dosage without any change in therapeutic response (Tr. at 933 (Dr. Barr)).^3a

3a Defendants agreed to this Finding "as a general principle."

58. Even where differences in bioavailability of products assumed to be equivalent do not result in clear risks to the patients, these differences can cause undesirable confusion in the drug marketplace (Tr. at 29 (Dr. Crout)).

V. The Need for In Vivo Bioavailability Testing, and the Standards by Which Such Tests Are Judged, Are Well Recognized by Scientists

A. Bioavailability Tests Can Substitute for More Extensive Clinical Testing

59. Prior to marketing of an innovator drug, the marketer must first show safety through animal studies, then progress to very limited controlled clinical studies to show efficacy, then more detailed clinical studies to show that the product will be safe and effective in controlled clinical settings in large numbers of people (Tr. at 887-88 (Dr. Barr)). (A)

60. When a second manufacturer seeks to market a copy of a drug product that has gone through the required clinical testing, it is "satisfactory from a scientific point of view" for that company to demonstrate that its drug product delivers that drug entity to the systemic circulation or the blood stream at exactly the same rate and the same amount as the product of the innovator, that is, by performing a well-conducted bioavailability study (Tr. at 891 (Dr. Barr)).

B. Bioavailability Testing, To Be an Acceptable Substitute for Clinical Tests, Must Meet Accepted Standards

61. There are three methods for determining bioavailability in vivo (in the human body): (1) determination of blood levels or urine levels of the drug or its metabolite following administration, the preferable method; (2) failing that, a valid pharmacological method may be acceptable; or (3) a clinical trial comparing clinical activity which is sensitive to the drug (Tr. at 476-78 (Dr. Cabana)). The three forms of testing are listed in descending order of accuracy, sensitivity, and reproducibility: the second method is acceptable only when the first is not possible and the third is acceptable only when the first and second are impossible to perform (21 C.F.R. § 320.24(c)). (A)

62. In vitro (in beaker) testing may be appropriate for determining the bioavailability only of drugs with no bioavailability problem or where in vitro tests have been correlated with the results of in vivo testing (Tr. at 478-81 (Dr. Cabana)). In vitro tests do not provide a basis for conclusions about bioavailability unless a correlation has been made between the results of those studies and the results of an in vivo study for the drug (Tr. at 898-99 (Dr. Barr); see also Tr. at 1615-16 (Dr. Rhodes); P. Ex. 169). In vitro tests cannot reliably establish the bioavailability of drug products when those products fall into any of the following three categories: (1) drugs at which dissolution does not appear to be related to bioavailability, for example, where the drug breaks down in the stomach; (2) drugs whose solubility is poor; and (3) drugs for which the minimum dissolution required to achieve good bioavailability is not known (Tr. at 481-83 (Dr. Cabana)). The three main types of bioavailability tests are serum level tests, bioassays, and in vitro tests correlated to in vivo testing (Tr. at 900-05 (Dr. Barr)). (A)

63. The usual method of determining whether products are bioequivalent is a blood level study, which is used to determine the relative rate and extent of absorption of the drug in the blood stream (Tr. at 898 (Dr. Barr)). In some cases one can use other body fluids, such as excretion of the drug into the urine, to assess the relative amount of the drug which has been absorbed (id.). The three usual parameters compared in a bioavailability blood level test are the area under the blood level curve (AUC), maximum concentration of the drug in the blood (C. Max.), and the time to reach the maximum concentration of the drug in the blood (T. Max.) (Tr. at 894-95 (Dr. Barr)).

64. The assay used in a bioavailability test is critical. It must be specific. It must measure only the drug and not drug metabolites, that is, those substances that the body forms after you give the drugs where the body transforms it into other materials which may be active or inactive. It must be reproducible "from one study to the next." It must be precise in its ability to measure exact amounts of the drug product (Tr. at 908 (Dr. Barr)).

65. Evaluation of a bioavailability study involves a testing of more than one hypothesis. First, one determines whether the test shows that with a 95% degree of confidence ( i.e., 95 times out of 100) the two products have not been shown to differ from each other (Tr. at 910-11 (Dr. Barr)).

Second, the "power" of the test is determined, that is, it is determined whether the test itself has enough precision so as to be able to detect the real differences between the drugs if differences exist, e.g., whether the test could detect a real difference of 20% between the bioavailabilities of the two drugs 80% of the time. In other words, if 100 studies of the design of that study were done, 80% of the time those studies would say that there was no difference greater than 20% between the two drugs (Tr. at 912 (Dr. Barr)). (NDBI)

C. A Bioavailability Test Must Have the Sensitivity, or "Power," To Detect a Difference if There Is One

66. The concept of "power" of a test used to determine bioequivalence is of tremendous significance in biopharmaceutics and pharmacokinetics (Tr. at 909 (Dr. Barr)). If one seeks, by performing a test, to determine whether two products differ from each other, a failure of that test to determine a difference may mean simply that the test is too insensitive to detect that difference rather than that there is no difference. For example, two aspirin tablets, one weighing 50% more than the other, put on a truck balance, would not be shown to weigh different amounts, simply because the instrument used is relatively insensitive to determine that difference. (Tr. at 909-10 (Dr. Barr)).

67. Mr. Robillard of Clinical Concepts would have done a power analysis on a final report of the tests to determine triamterene blood levels and hydrochlorothiazide blood levels in a bioavailability study of Premo's Triamthiazide product had that study been completed. (P. Ex. 84-A, page 2; P. Ex. 84-B at 328 (Deposition of N. Robillard). (NDBI)

68. Both Dr. Westlake and Dr. Wagner, relied upon by Dr. Greene, in books they authored set out formulas for power analyses (Tr. at 1915-17 (Dr. Greene); P. Ex. 250). The alternative theory to power analysis, developed by Dr. Westlake, an approach to variance based on confidence intervals, was not used by Dr. Greene in analyzing the Triamthiazide or chlorthalidone bioavailability studies (D. Ex. 83 revised, 84 and 39-A) (Tr. at 1917-18 (Dr. Greene)).

69. While Dr. Greene testified that he found only three articles in the Journal of Pharmaceutical Sciences between 1970 and 1980 that involved bioavailability tests with power analysis, he could recall few articles concerning bioavailability tests published in that journal, which limits publication to "state of the art developments" (Tr. at 1891 (Dr. Greene)). (NDBI)

70. A power analysis is absolutely necessary for a complete bioavailability test, since otherwise a test involving a great deal of variability among subjects or with an insensitive assay could be used to show that there was no detected difference between two drug products (Tr. at 913 (Dr. Barr)). A bioavailability test without some way to evaluate the overall sensitivity of the test is meaningless (Tr. at 914).

VI. The Need for Premarket Clearance of Generic Copies Is Recognized by Those Knowledgeable in the Field

71. It is necessary to be concerned about the quality of drug products very much more than other consumer products because if drug products fail in quality the results can be extremely serious, disastrous in some cases (Tr. at 1413 (Dr. Rhodes)). A definition for quality for drug products includes release of the active ingredient for a full biological effect (Tr. at 1414 (Dr. Rhodes)). (NDBI)

72. It is reasonable to require that a new drug product be demonstrated to be safe and effective (Tr. at 1416-17). (NDBI)

73. If there were no preclearance system in the form of approved new drug applications or abbreviated new drug applications, drug companies could develop products containing the same active ingredients as products already marketed, use safe and suitable active and inactive ingredients, follow good manufacturing practices and comply with all compendium standards in existence and yet produce a product with significantly different bioavailability from the product already on the market (Tr. at 1487-93 (Dr. Rhodes)). If there were no such preclearance mechanism it would be possible that such a biological availability difference would not be known until the product was actually physically in the market place (Tr. at 1493-94 (Dr. Rhodes)). (A)

74. The chance of companies manufacturing products having differences in bioavailability is too great simply to allow those products onto the market without someone reviewing them first, even if the same active or same amount of active drug ingredient is present (Tr. at 831 (Dr. Shangraw)).

75. At the present time, the Food and Drug Administration is the only "reasonable body" capable of reviewing drug products for bioequivalence (Tr. at 871 (Dr. Shangraw)).

76. An incident in this proceeding demonstrated the necessity of a review by FDA experts, not employed by the sponsoring drug company, of testing done to support that company's drugs. Dr. Rhodes, one of defendant's experts, prior to listening to Dr. Cabana's "penetrating analysis" of the Premo chlorthalidone study, was unaware of problems with that study and made conclusions that "had to be re-examined" after the review by the FDA and by Dr. Cabana pointed out problems with the study (Tr. at 1693-94 (Dr. Rhodes)). Dr. Rhodes agreed that when he stated in his sworn affidavit that "Premo has commissioned a well-designed and executed bioavailability study in which their product was compared with the bioavailability of the innovator's product," he was wrong (Tr. at 1694 (Dr. Rhodes)).

77. If FDA must, as part of enforcement actions, review, for court evaluation, studies concerning drug products designed to show those products' actual safety and effectiveness each time FDA seeks to remove a new drug product from the market pending premarket review, "the FDA will very soon be unable to perform with any degree of fairness to other manufacturers its assigned task of reviewing legally submitted NDAs and ANDAs" (P. Ex. 5 (Aff. of Dr. Crout), ¶ 14, page 7).

78. The Food and Drug Administration has developed a compliance policy with respect to enforcement of the "new drug" provisions as they relate to prescription drugs that constitutes a rational determination of how to use limited enforcement resources to effectuate compliance with the law (Tr. at 67, 70-79, 80 (Dr. Crout); P. Ex. 4 (Aff. of Dr. Crout), ¶¶ 4-8). The Food and Drug Administration's compliance policy has been made widely available to the drug industry (Tr. at 81-82 (Dr. Crout)), and can not be said to constitute arbitrary decisionmaking by the agency. (A)

VII. Compliance with Compendial Guidelines Is No Substitute for Premarket Clearance

79. The United States Pharmacopeia ("U.S.P.") does not assume responsibility for insuring bioequivalence of different drug products for a particular drug and compliance with U.S.P. monographs does not assure bioequivalence (Tr. at 936-37 (Dr. Barr); Tr. at 726 (Dr. Shangraw)). Tetracycline capsules, all of which met U.S.P. standards, have been shown to produce significant differences in bioavailability (Tr. at 1011 (Dr. Barr)). (NDBI)

80. The U.S.P. and the National Formulary ("N.F.") both lack the internal and external resources required to develop a rational and modern standard. The reliance on such limited resources is anachronistic and inadequate for the scientific activities necessary to ensure pharmaceutical equivalence of drug products (Tr. at 1520 (Dr. Rhodes)). (NDBI)

81. Many active ingredients and excipients are not included in the U.S.P. even today (Tr. at 1520 (Dr. Rhodes)). There is, in any case, no monograph in an official compendium for a drug product containing the combination of triamterene and hydrochlorothiazide (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 35.a), nor is there anywhere else a publicly available monograph for this drug to establish a standard for identity, quality, strength, and purity for drug products containing these ingredients (id., ¶ 35.b). (A)

82. Even where excipients are included in a U.S.P. monograph, specific tests and standards applicable to their use as components in drug products such as tablets or capsules are not always available (Tr. at 1520-21 (Dr. Rhodes)). (A)

83. At least with respect to some drugs in the U.S.P. the monographs do not address the problem of interaction between active ingredients and excipients (Tr. at 1521 (Dr. Rhodes)). (A)

84. Most U.S.P. monographs do not have specifications for granulation or precomposition mixtures (Tr. at 1521 (Dr. Rhodes)), even though with respect to some drugs the particle size of the active ingredient is quite important (Tr. at 1522). (A)

85. Only a small number (10-12) of the drugs in the new combination National Formulary and United States Pharmacopeia have in vitro or dissolution monographs that are correlated with in vivo results (Tr. at 728 (Dr. Shangraw)). In the absence of a correlation between in vitro dissolution studies and in vivo bioavailability studies there can be no assurance that a dissolution test will demonstrate bioequivalence of one drug to another (Tr. at 727-28, 729-30 (Dr. Shangraw)). (A)

86. U.S.P. disintegration tests are so imprecise as to have a 40% chance of allowing defective tablets to pass the test (Tr. at 1523-25 (Dr. Rhodes)). Some dissolution testing in the U.S.P. is so imprecise as to allow a 60% error to go undetected (Tr. at 1524-27 (Dr. Rhodes)).

87. If there are polymorphic differences in the active ingredient of a drug, that type of difference would not be detectable by methods included in the U.S.P. monographs (Tr. at 571-72 (Dr. Cabana)).

88. Different methods of manufacturing active ingredients may cause different degradation products (Tr. at 1425 (Dr. Rhodes)). The compendial, i.e., U.S.P.-N.F., assay standards assume that certain processes of drug synthesis have been used, and therefore it is conceivable that the prescribed assay may not pick up other degradation products and other impurities in an active drug substance synthesized in a different manner (Tr. at 1420-21 (Dr. Rhodes)). (NDBI)

89. A U.S.P. monograph would not be changed until there was clear and convincing evidence of the inadequacies of the assay in that monograph, although if a serious problem came to light in a monograph this could be changed (Tr. at 1423-23 (Dr. Rhodes)).

90. Even though digoxin in tablet form can vary greatly with respect to bioavailability (Tr. at 1511 (Dr. Rhodes)), and people began to realize that fact about 15 years ago ( id.), there was a monograph in the U.S.P. in 1970, compliance with which did not insure that the drug product containing digoxin did not have significant bioavailability problems (Tr. at 1512-13; P. Ex. 212). (NDBI)

91. Dr. William Barr, one of the six trustees who serve as a policymaking group of the U.S.P. (Tr. at 878-79 (Dr. Barr)), and Dr. Shangraw, a member of the executive committee of the U.S.P. and chairman of the Formulations Committee of the Committee of Revision of the U.S.P. (P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 7; Tr. at 713 (Dr. Shangraw)), believe FDA premarketing clearance of generic copies of approved drugs is essential ( see P. Ex. 1 (Aff. of Dr. Barr), ¶¶ 19-209; P. Ex. 7 (Aff. of Dr. Shangraw), ¶ 17; Tr. at 831 (Dr. Shangraw)). (NDBI)

VIII. Compliance with Good Manufacturing Practice Regulations Does Not Adequately Assure the Safety and Effectiveness of Generic Copies of Approved Drugs

92. Compliance with good manufacturing practices ("GMPs") would not assure the safety and effectiveness of generic copies of innovator products for the following reasons:

A. GMPs do not specify which excipients should be used in any particular drug product (Tr. at 1527-28 (Dr. Rhodes)).

B. GMPs do not specify particular manufacturing processes (Tr. at 1528).

C. GMPs do not specify the source of the active ingredient of the drug product ( id.).

D. GMPs do not specify the particle size with respect to manufacturing processes ( id.).

E. GMPs do not guarantee that any particular product will be stable (Tr. at 1529). (NDBI)

93. GMP inspections by FDA inspectors would not be a substitute for premarket clearance because FDA inspectors are not experts in pharmacokinetics or biopharmaceutics (Tr. at 1530 (Dr. Rhodes)), only a sampling of drug products are selected for inspection during a GMP inspection (Tr. at 1531), most FDA inspectors are not specialists in drug formulation and do not normally have expertise in determining what excipients would be used in formulating particular drug products ( id.), and most FDA inspectors have no particular expertise as to techniques used in manufacturing ( id.; see also Tr. at 1532). Also, it is possible for a drug manufacturer to be manufacturing and distributing a drug product for up to six months before the FDA is even aware that the manufacturer is selling the product (Tr. at 1532 (Dr. Rhodes)), and GMP inspections can be as infrequent as once every other year ( id., see also Tr. at 319-20 (Dr. Crout)). (NDBI)

94. When Mr. John Blackman stated that there were at least 85 visits by Food and Drug Administration inspectors as part of good manufacturing practice inspections at the Premo plant during a certain recent period of time, he was including inspections other than GMP inspections, i.e., inspections to learn of possible violations of the new-drug provisions of the Federal Food, Drug and Cosmetic Act (P. Ex. 81-A, page 1; P. Ex. 81-B at 10 (Deposition of J. Blackman)).⁴ (NDBI)

4 The deposition for cross-examination of John Blackman was numbered separately from the deposition containing his direct examination. These numbers refer to the cross-examination on June 16, 1980.

IX. Premo's Actions Demonstrate That if There is No Premarketing Clearance Requirement for Generic Copies of Approved Drugs, Generic Manufacturers Will Place Drugs on the Market Without Proper Testing and Without Even Consulting Their Expert Consultants

A. Marketing Without Testing

95. Premo began marketing chlorthalidone tablets in June 1979 (P. Ex. 64). It has presented no evidence of any testing of this product that was completed at that time. The flawed bioavailability test of chlorthalidone was not analyzed and reported to Premo until December of 1979 (D. Ex. 39; P. Ex. 84-A, page 1; P. Ex. 84-B at 150 (Deposition of N. Robillard)). No evidence has been presented that Premo even sought to analyze the innovator product to determine its ingredients so that it could be copied. (NDBI)

96. Premo began to market allopurinol tablets in October 1979 (P. Ex. 64), despite the fact that no in vivo bioavailability testing of this product had been completed at that time. No evidence has been submitted of an analysis of the innovator product to determine whether Premo's product was equivalent to it in ingredients that predated the marketing of Premo's allopurinol. Fittelson Laboratories, with which United States Testing Company contracts to analyze drug products, submitted a report on the innovator allopurinol product that dealt only with analysis to determine the amount of lactose and the amount of starch in the product on November 8, 1979 (P. Ex. 83-A, page 1; P. Ex. 83-B at 61-63 (Deposition of S. Kahan); D. Ex. 306-1). (NDBI)

97. Premo began marketing betamethasone valerate cream in May of 1979, (P. Ex. 64). The record reveals no testing of that product that had been completed at that time. Nor is there any evidence in the record that Premo had analyzed the innovator product to determine whether its product contained the same active ingredients as the innovator. (NDBI)

98. Premo began to market its hydroxyzine hydrochloride tablets in April 1978 (P. Ex. 64). There is no evidence in the record that there was any testing of this product before marketing or that Premo sought to analyze the innovator product to match its ingredients. (NDBI)

99. Premo began to market its hydroxyzine pamoate capsules in May of 1978 (P. Ex. 64). There is no evidence in the record that Premo performed any bioavailability or other testing on this product before marketing, and no evidence that it analyzed the innovator product to determine its ingredients to copy them prior to marketing its product. (NDBI)

100. Premo began to market its product containing doxylamine succinate and pyridoxine hydrochloride in September of 1978 (P. Ex. 64). Premo presented no evidence that it had performed any bioavailability or other testing of this product prior to marketing, or that it had analyzed the innovator product in an attempt to match its active ingredients. (NDBI)

101. Premo began to market its trifluoperazine hydrochloride tablets in November of 1979 (P. Ex. 64). Premo has presented no evidence that it performed bioavailability or other testing of these tablets prior to marketing or that it analyzed the innovator product in an attempt to match its ingredients prior to the marketing date. (NDBI)

102. Premo began to market its Triamthiazide (triamterene plus hydrochlorothiazide) capsules on November 14, 1979. (NDBI)

103. Premo had not received any bioavailability or blood level data with respect to the Triamthiazide product prior to marketing (Tr. at 2238 (S. Blackman)). (NDBI)

104. Premo had no information before it started marketing Triamthiazide that its product was bioequivalent to the innovator Dyazide (P. Ex. 25-A, page 4; P. Ex. 25-B at 54 (Deposition of J. Blackman)). (NDBI)

105. When Premo began to market its Triamthiazide product it had no clinical data with respect to the edema indications for that product (P. Ex. 26-A, page 7; P. Ex. 26-B at 203 (Deposition of J. Blackman); Tr. at 2238 (S. Blackman)). (NDBI)

106. Mr. Seymour Blackman states he was present at a phone conversation in which an oral final report of a clinical study concerning a Triamthiazide product was received by Walter Epler from Norman Robillard prior to the decision to market Premo's Triamthiazide product (Tr. at 2237 (S. Blackman)); Walter Epler, however, denied having received an oral report concerning the values or conclusions to be drawn from the clinical study concerning Premo's Triamthiazide product before receipt of the final report dated November 26, 1979 (P. Ex. 27-A, page 3; P. Ex. 27-B at 25). (NDBI)

107. While John Blackman states that Premo relies upon its experts, Premo required Dr. Rhodes' advice to do a bioavailability study of Triamthiazide (Tr. at 1430-31 (Dr. Rhodes)). As to Dr. Tessler, upon whom Premo also said it relied, and who John Blackman thought was monitoring the clinical study of Triamthiazide, he was, as Blackman knew, in federal court as a party during the period from September and October up until November of 1979 (P. Ex. 25-A, page 4; P. Ex. 25-B at 48-49, 50 (Deposition of J. Blackman)).

108. John Blackman has admitted that Premo received the final report of the clinical study of Premo's Triamthiazide product on November 26, twelve days after marketing was begun (P. Ex. at 25-A, page 3; P. Ex. 25-B at 47). (NDBI)

109. Premo had received an interim report of the clinical study of Triamthiazide from Dr. Tessler prior to marketing the product (P. Ex. 25-A, pages 3-4; P. Ex. 25-B and 47-48 (Deposition of J. Blackman)). (NDBI)

110. The interim report, dated August 21, 1979, received by Premo prior to its marketing of its Triamthiazide product, states that Premo's Triamthiazide "is equally, if not more, effective than Dyazide in the control of uncomplicated essential hypertension" (emphasis added) (P. Ex. 26-C-6A, page 3). (NDBI)

111. Premo does not know the manufacturing techniques used by Smith, Kline and French in its production of Dyazide and Premo is not 100% positive about the actual ingredients used in the manufacture of Dyazide (P. Ex. 26-A, page 6; P. Ex. 26-B at 186-87). (NDBI)

B. Marketing Without Consulting Experts

112. Mr. John Blackman states that Dr. Rhodes consults with his department on a regular basis (P. Ex. 25-A, page 2; P. Ex. 25-B at 17; cf. Tr. at 1427 (Dr. Rhodes)). (NDBI)

113. Neither Dr. Rhodes nor Dr. Greene was consulted on whether Premo should market its Triamthiazide product and neither knew that it was being marketed until one of their graduate students brought each a sample capsule which had been obtained on the open market (Tr. at 1326, 1442 (Dr. Rhodes); Tr. at 1928-29 (Dr. Greene)). (NDBI)

114. Dr. Rhodes was not consulted with respect to the marketing of Premo's trifluoperazine hydrochloride product and did not learn that it was being marketed until he saw that it was named as one of the drugs in this case (Tr. at 1443 (Dr. Rhodes)). Dr. Greene did not even know at the time of his testimony that Premo was selling trifluoperazine (Tr. at 1931-32 (Dr. Greene)). (A)

115. Dr. Rhodes was not consulted prior to Premo marketing its chlorthalidone products and did not learn of that marketing until this case began (Tr. at 1444 (Dr. Rhodes)). Dr. Greene was not consulted on the marketing of chlorthalidone and learned about it when a pharmacy student brought a tablet to him (Tr. at 1929-30 (Dr. Greene)). (NDBI)

116. Dr. Rhodes was not consulted with respect to the marketing of Premo's allopurinol product and did not know that it was being marketed prior to the start of this case (Tr. at 1445 (Dr. Rhodes)). Dr. Greene was not aware even at the time of the law suit that Premo was marketing allopurinol (Tr. at 1932 (Dr. Greene)). (NDBI)

117. Dr. Rhodes did not become aware that Premo was marketing betamethasone valerate cream until August or September of 1979 and at that time did not know whether or not Premo had submitted a new drug application for the product (Tr. at 1445-46 (Dr. Rhodes)). Dr. Greene was not consulted on whether this product should be marketed (Tr. at 1933 (Dr. Greene)). (NDBI)

118. Prior to this case Dr. Rhodes was not aware that Premo was marketing its doxylamine succinate with pyridoxine hydrochloride combination (Tr. at 1446 (Dr. Rhodes)). Dr. Greene was not consulted on the marketing of this drug (Tr. at 1933 (Dr. Greene)). (NDBI)

119. Prior to this case, Dr. Rhodes was not aware that Premo was marketing either its hydroxyzine hydrochloride or hydroxyzine pamoate products (Tr. at 1447 (Dr. Rhodes)). Dr. Greene did not know at the time of trial whether hydroxyzine hydrochloride or hydroxyzine pamoate were being marketed by Premo (Tr. at 1932 (Dr. Greene)). (NDBI)

X. Dr. Rhodes' Suggestion That Bioequivalent Generic Products Can Be Made by Analysis of Approved Products in an Attempt To Copy Their Ingredients, a Suggestion Not Followed by Premo in Most Instances in any Case, Is Not Supported by Evidence in the Record

120. If a drug company starts out to make a generic version of a product and uses different generic components and different methods of preparing the drug product, that approach is likely to produce a product that may show significant biological availability differences between the generic product and the innovator product (Tr. at 1041-42 (Dr. Rhodes)). (A)

121. An alternative approach, acceptable in Dr. Rhodes' view, to formulating a drug product includes making an attempt, through use of an independent testing laboratory, to determine the constituents of a drug product (Tr. at 1494, 1042 (Dr. Rhodes)). Yet, there was no evidence that Premo followed this procedure for any of its products other than Triamthiazide prior to marketing or for any product other than allopurinol after marketing ( see Proposed Findings 98-104, supra). (NDBI)

122. Fittelson Laboratories, which tested Premo products by contract with United States Testing Co. (P. Ex. 83-A, page 2; P.Ex. 83-B at 101 (Deposition of S. Kahan)), is seldom asked to break down a tablet or capsule into the excipients and lubricants and other materials which may be present (P. Ex. 83-A, page 1; P. Ex. 83-B at 10-11 (Deposition of S. Kahan)). That laboratory, while equipped to do so, would be reluctant to do a complete analysis of a tablet without an idea of the individual components of the tablet (P. Ex. 83-A, page 1; P. Ex. 83-B at 11-12 (Deposition of S. Kahan); see also Tr. at 1497 (Dr. Rhodes)).

Paragraphs 123 through 128 are contained in the sealed portion of these findings.

XI. Premo's Products Differ from the Products for Which They May Be Substituted and Those Differences May Be Important

A. Premo's Triamthiazide Product Differs Significantly in Composition from the Innovator Product and Has Not Been Shown To Be Bioequivalent or Equivalent in Safety and Effectiveness to the Innovator Product for Which It May Be Substituted

1. If Premo's Triamthiazide Differs in Bioavailability from the Innovator Product, That Fact Could Cause Serious Harm to Patients

129. The use of a Triamthiazide product that is more bioavailable than Dyazide could lead to a significant danger of hyperkalemia which can cause the heart muscle to beat irregularly and stop, thus leading to death (P. Ex. 3 (Aff. of Dr. Cortell), ¶ 19; P. Ex. 4 (Aff. of Dr. Crout), ¶¶ 15-17; P. Ex. 2 (Aff. of Dr. Cabana), ¶ 39.c). Use of Triamthiazide product that is significantly more bioavailable than Dyazide could also lead to the formation of kidney stones (renal calculi) (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 39.d).

130. If the Premo Triamthiazide product were less bioavailable than the Dyazide product that fact would also lead to risk to patients because patients treated with the Triamthiazide would find that drug ineffective in controlling the conditions for which it was prescribed, a therapeutic failure that can lead to, among other things, dysfunction of cardiac rhythm and sudden death (P. Ex. 3 (Aff. of Dr. Cortell), ¶ 20; P. Ex. 4 (Aff. of Dr. Crout), ¶¶ 12 and 14; P. Ex. 2 (Aff. of Dr. Cabana), ¶ 39.e).

131. Patients taking drugs such as triamterene plus hydrochlorothiazide are often on other drug products as well (Tr. at 1575 (Dr. Rhodes)), and often have other diseased states which are being treated by these other drugs ( id.), such as instances in which drugs such as Digitalis would be used ( id.). This fact can increase the medical significance of any bioavailability problem with Triamthiazide (Tr. at 618-19 (Dr. Cabana)).

132. Many patients taking a combination product containing triamterene with hydrochlorothiazide have a condition known as diabetes mellitus, adult onset (Tr. at 1575 (Dr. Rhodes)) and with respect to such patients an increasingly potent triamterene with hydrochlorothiazide would present distinct possibility of complications ( id.). (NDBI)

133. The toxic effects associated with combinations of triamterene plus hydrochlorothiazide represent a relatively steep dosage response, that is, an increase in effective dosage of the drug brings a marked increase in the toxic effects for the drug (Tr. at 934-36 (Dr. Barr)).

134. In one study involving elderly patients, twenty-two patients treated with triamterene plus hydrochlorothiazide moved from a normal serum potassium range into the range defined by Dr. Rhodes as hyperkalemic (Tr. at 1585 (Dr. Rhodes); D. Ex. 140, page 170, table 5).

135. Hypokalemia is dose responsive to hydrochlorothiazide, that is, an increase in the amount of hydrochlorothiazide is likely to bring an increase in the potassium excretion of the patient (Tr. at 1622-24 (Dr. Rhodes)). (NDBI)

2. There Are Bioavailability Problems with Products That Contain, as Does Triamthiazide, Triamterene Plus Hydrochlorothiazide

136. Dr. Rhodes advised Premo it was necessary to perform a bioavailability study for the drug product containing triamterene and hydrochlorothiazide even though they might have the clinical study underway and he told them it was perfectly feasible to assay the drug in body fluids (Tr. at 1430-31 (Dr. Rhodes)). (NDBI)

137. FDA's files contain data which identify significant differences in both the dissolution and bioavailability of products containing triamterene plus hydrochlorothiazide (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 36).

138. Differences in bioavailability of triamterene plus hydrochlorothiazide products may come from changes in manufacturing processes, such as simply changing from a laboratory batch to a full production batch or from changing the order in which the components of the drug are mixed ( id., ¶ 37.b).

139. A bioavailability study in the files of the FDA has shown a difference in peak plasma levels of four to five times between a formulation containing triamterene plus hydrochlorothiazide and the currently marketed product, Dyazide ( id., ¶ 37.c). (NDBI)

140. A protocol submitted by Premo to the FDA for its product Triamthiazide states that preliminary study, which Premo now argues was of an experimental formula, suggests a seven to eight fold difference in bioavailability of the Premo formulation compared to the marketed product, Dyazide ( id., ¶ 38.c). This is not, however, entitled to substantial weight, since it represents data generated from only one subject taking the Premo formulation in comparison with an entirely different subject taking the Dyazide formulation.

141. Innovator products may not be well absorbed so that products that imitate the innovator's may be absorbed a great deal more than the innovator's. Thus, for example, if only 10% of the innovator product is absorbed in the blood the imitator may be absorbed as much as 10 times as much as the innovator (Tr. at 977-78 (Dr. Barr)). The innovator product copied by Triamthiazide is probably absorbed only 10 to 20% itself. (Tr. at 977).

142. Dr. Rhodes, in a February 14, 1980, evaluation of a preliminary bioavailability study of Premo's triamterene with hydrochlorothiazide products saw bioavailability problems with the drug, stating that the data in that study "does not indicate bioequivalence, rather it suggests that the Premo products are yielding a significantly higher proportion of triamterene. I suspect that this finding indicates some major problem in the Smith Kline product." (P. Ex. 244, page 2; Tr. at 1652 (Dr. Rhodes)). Dr. Rhodes, at trial, attempted, not entirely successfully, to explain away this evaluation.

143. The two active ingredients of drugs containing triamterene and hydrochlorothiazide present bioavailability problems. See ¶¶ 144-147, infra.

144. Triamterene is a drug with a known bioavailability problem (Tr. at 1597 (Dr. Rhodes)), as Dr. Rhodes advised Premo on September 21, 1979 ( id.). Triamterene is extremely insoluble in water, a fact that suggests a possible bioavailability problem ( id.; Tr. at 484 (Dr. Cabana); Tr. at 1895, 1900 (Dr. Greene)). Triamterene is subject to considerable metabolism and is subject to a first pass effect that may alter the absorption of the drug into the human system (Tr. at 1598 (Dr. Rhodes); Tr. at 1897 (Dr. Greene)). (NDBI)

145. There has been documentation of up to a 300% difference in bioavailability of triamterene by two different firms in submissions to the FDA (Tr. at 484, 590 (Dr. Cabana)).

146. There is documentation in the published literature and in studies submitted to the FDA that the drug hydrochlorothiazide has a bioavailability problem (Tr. at 484, 589-91, 616 (Dr. Cabana)).

147. Hydrochlorothiazide is a product with limited solubility (Tr. at 484 (Dr. Cabana)). (NDBI)

3. The Formulation of Premo's Triamthiazide, When Compared to the Formulation of the Innovator Dyazide, Suggests That Triamthiazide Will Differ from Dyazide in Bioavailability

Paragraphs 148 through 151 are contained in the sealed portion of these findings.

152. There are, based upon their formulations, predictable differences in bioavailability between the Premo Triamthiazide drug and the innovator product it imitates (Tr. at 833 (Dr. Shangraw)).

B. Premo's Trifluoperazine Product Differs Significantly in Composition from the Innovator Product and Has Not Been Shown To Be Bioequivalent or Equivalent in Safety and Effectiveness to the Innovator Product for Which It May Be Substituted

1. If There Is a Difference Between the Bioavailability of Premo's Trifluoperazine and the Innovator Product, That Fact Can Cause Serious Harm to Patients

153. Drug products containing trifluoperazine hydrochloride ("trifluoperazine") are given in small amounts (Tr. at 335 (Dr. Perel)). Trifluoperazine is an extremely potent drug, which is used in both children and adults ( id.). (A)

154. Trifluoperazine products have a narrow therapeutic range, i.e., the dosage in the blood at which one observes adverse effects is very close to the dosage in the blood at which the chief therapeutic response is observed (Tr. at 336 (Dr. Perel); P. Ex. 69-A, page 1; P. Ex. 69-B at 78 (Deposition of Dr. Seife)). (A)

155. Trifluoperazine causes several types of serious side effects: pseudo-parkinsonism side effects, i.e., trembling or shaking, painful muscular spasms, etc., an effect occurring at low blood levels of the drug (Tr. at 337 (Dr. Perel)); tardive dyskinesia, which takes place after chronic administration of trifluoperazine and involves involuntary movements affecting the mouth and trunk and extremities and which can cause permanent irreversible impairment in 10-20% of those persons exhibiting those side effects (P. Ex. 6 (Aff. of Dr. Perel), ¶ 10; Tr. at 338-40 (Dr. Perel); see also P. Ex. 69-A, page 1; P. Ex. 69-B at 78 (Deposition of Dr. Seife)); and cardiovascular side effects, including a significant amount of hypertension (Tr. at 339 (Dr. Perel)). (A)

156. If a clinician used a generic trifluoperazine product that was more bioavailable than Stelazine, the approved product containing trifluoperazine, that would increase the possibility of tardive dyskinesia. This is true because the effects of the product which was more bioavailable than expected would be masked since the clinician would prescribe a higher dosage to cover the first signs of involuntary movements associated with tardive dyskinesia (Tr. at 340-43 (Dr. Perel)). This effect is more likely to occur with institutionalized patients who would not be observed as closely by physicians as would those that are not institutionalized (Tr. at 347-48 (Dr. Pere])). (A)

157. If a clinician prescribes a generic drug purportedly containing the same active ingredient as Stelazine but less bioavailable than Stelazine, that drug would not have the expected therapeutic effect. The fact that the drug was not working would not be discovered for from 8 to 14 weeks, during which time the patient would be institutionalized. (Tr. at 344 (Dr. Perel)). (A)

158. If a patient is stabilized on Stelazine, and a drug is substituted for Stelazine that is less bioavailable, the patient would relapse but the relapse would likely be considered disease-related rather than drug-related (Tr. at 344-45 (Dr. Perel)). (A)

159. If lower bioavailability of a generic product leads a clinician to conclude that trifluoperazine is ineffective in a patient, that patient may simply not be helped by any medication, may be unable to function in the community, and may thus be institutionalized (Tr. at 345-47 (Dr. Perel)). (A)

160. The committee on Bioavailability and Bioequivalence of Psychotropic Drugs of The American College of Neuropsychopharmacology has recommended against use of multi-source psychotropic drug products such as trifluoperazine interchangeably unless bioequivalence has been shown (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 42.g). (A)

161. Bioequivalence problems with trifluoperazine would present a medically significant problem that might go unrecognized in the context of treatment of patients with this drug (Tr. at 620-21 (Dr. Cabana)). (A)

2. Trifluoperazine Is a Drug with Potential Bioavailability Problems

162. Trifluoperazine hydrochloride ("trifluoperazine") is a member of the phenothiazine drug class for which there are documented bioavailability problems (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 42.c; Tr. at 489, 613-14 (Dr. Cabana)). (A)

163. Trifluoperazine is a drug whose rate and extent of absorption, i.e., bioavailability, is extremely critical because the drug is rapidly metabolized (changed) by the liver. Thus, if the rate of absorption is slow, the first pass metabolism by the liver can obliterate the pharmacological effects of the drug because inadequate amounts of the drug will be available to act at the site of action. (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 42.d; Tr. at 489, 614 (Dr. Cabana)). Alternatively, if the imitator drug has too fast absorption, it may send much more drug to the site of action because the liver would break down less of the drug during the first pass than it would of a less quickly absorbed drug (Tr. at 818-19 (Dr. Shangraw)). (A)

3. Comparison of the Formula of Premo's Trifluoperazine and That of the Innovator Product Suggests That the Two Products Will Differ in Bioavailability

164. The ingredients in Premo's trifluoperazine products differ as to inactive ingredients from those of the innovator products (P. Ex. 12; P. Ex. 15). (A)

Paragraphs 165 through 167 are contained in the sealed portion of these findings.

4. Although Bioavailability Testing of Trifluoperazine Is Possible, Premo Has Done None

168. There are, based on their formulations, predictable differences in bioavailability between the Premo trifluoperazine product and the innovator it imitates (Tr. at 833 (Dr. Shangraw)). (A)

169. There are three known means of assaying trifluoperazine in the blood (Tr. at 348-56 (Dr. Perel); Tr. at 545 (Dr. Cabana)) — radioimmunoassay techniques (Tr. at 350, 352 (Dr. Perel)); Tr. at 545 (Dr. Cabana)); radio receptor technique (Tr. at 352-353 (Dr. Perel); Tr. at 545 (Dr. Cabana)); and GCMS (Gas Chromotography-Mass. Spectrometry) (Tr. at 354-55 (Dr. Perel); Tr. at 545-46 (Dr. Cabana)).

170. The method development study using measures of pupillometry to attempt to quantitate amounts of trifluoperazine did not measure any Premo product and is not a bioavailability study that would show Premo's product to be bioequivalent to the approved product (Tr. at 368-69 (Dr. Perel)). (NDBI)

Premo has presented no evidence that any other bioavailability test of its trifluoperazine product was performed.

C. Premo's Chlorthalidone Product Differs Significantly in Composition from the Innovator Product and Has Not Been Shown To Be Bioequivalent or Equivalent in Safety and Effectiveness to the Innovator Product for Which It May Be Substituted

1. If Premo's Chlorthalidone Product Differs in Bioavailability from the Innovator Product, That Fact Could Cause Serious Harm to Patients

171. If Premo's chlorthalidone is more bioavailable than the marketed product it can produce unwanted toxic effects, including hypokalemia which, particularly when this drug is used concurrently with other drugs, can lead to serious cardiac toxicity and possibly sudden death (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 41.e).

172. If Premo's chlorthalidone product is significantly less bioavailable than the approved product, Premo's drug would be inadequate to control high blood pressure in edema for which it is prescribed, leading to therapeutic failure which can cause risk to a patient. (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 41.e; P. Ex. 3 (Aff. of Dr. Cortell), ¶ 22; P. Ex. 4 (Aff. of Dr. Crout), ¶¶ 13-14).

173. The medical significance of bioavailability problems with chlorthalidone will be enhanced in patients treated for heart disease and cardiac conditions who might be treated concurrently with other drugs (Tr. at 618-19 (Dr. Cabana)).

174. A switch from 50 mg of chlorthalidone to 75 or 100 mg of chlorthalidone would alter serum potassium levels, an effect particularly dangerous in patients on digitalis (Tr. at 603 (Dr. Cabana); Tr. at 985 (Dr. Barr)). (NDBI)

175. A bioavailability difference of less than two-fold, 100% increase, or 100% decrease, would be definitely significant in chlorthalidone in terms of the effect on potassium (Tr. at 618 (Dr. Cabana)).

176. The dose response curve for the toxicity of chlorthalidone is steep, that is, increases in the effective dosage of chlorthalidone result in increased uric acid levels and increased potassium excretion at a relatively high rate (Tr. at 934-35 (Dr. Barr); P. Ex. 63). The article by Tweedale et al. shows clear dose response curves that are steep for serum potassium, which would correlate with potassium excretion (Tr. at 986-87).

177. The incidence of hypokalemia, hyperuricemia, and increased serum chloride, all potentially toxic effects, increases at a small rate as dosages of chlorthalidone increase (Tr. at 1629 (Dr. Rhodes)). There is no plateau curve in the dose response to these products ( id.).

178. The dosage of chlorthalidone that would be toxic would depend upon the individual involved (Tr. at 574 (Dr. Cabana)). (NDBI)

179. According to data submitted to FDA, an increase in dosage from 50 mg chlorthalidone to 75 mg chlorthalidone has been shown to be toxic in some patients (Tr. at 574-75 (Dr. Cabana)). Dosages above 50 mg of chlorthalidone cause a change in potassium in a very large percentage of subjects. That change may be dangerous if the patient is taking digoxin. Two million patients in the United States do take digoxin, and many of those patients are also taking chlorthalidone (Tr. at 575 (Dr. Cabana)).

180. Dr. Rhodes is aware of no clear statement in the literature that there is no medically significant bioavailability problem with chlorthalidone (Tr. at 1220-21 (Dr. Rhodes)). (A)

181. An explanation for why there has been no documented therapeutic failure due to the bioavailability of chlorthalidone is that there has been no product other than the approved product on the market until very recently (Tr. at 605 (Dr. Cabana)).

182. A hundred milligram chlorthalidone dose is not justified in view of the toxic side effects which have been documented (Tr. at 1633 (Dr. Rhodes); see also P. Ex. 4, Ex. 6). (NDBI)

2. Chlorthalidone Is a Drug with Known Bioavailability Problems

183. Chlorthalidone has known bioavailability problems (Tr. at 1629-30 (Dr. Rhodes)).

184. Drug products containing chlorthalidone present a bioavailability problem because chlorthalidone has a solubility problem (Tr. at 485, 602 (Dr. Cabana); P. Ex. 4 (Aff. of Dr. Crout), Attachment 6). (A)

185. There is no U.S.P. dissolution standard appropriate to assure the bioavailability of chlorthalidone (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 41.a; Tr. at 566-68; D. Ex. 82). (NDBI)

3. Comparison of the Formulations of Premo's Chlorthalidone Product and That of the Innovator Product Suggests Potential Differences in Bioavailability Between the Premo Product and the Innovator Product

186. The ingredients in Premo's chlorthalidone products differ, in some respects, from those in the innovator products (P. Ex. 12; P. Ex. 13). (NDBI)

Paragraphs 187 through 188 are contained in the sealed portion of these findings.

189. The differences between the USV chlorthalidone product and the Premo chlorthalidone product could significantly affect bioavailability (Tr. at 813 (Dr. Shangraw)), that is, in view of the differences one could not with any reasonable degree of certainty predict bioequivalence for the drug (Tr. at 814).

D. Premo's Allopurinol Product May Differ Significantly in Composition from the Innovator Product and Has Not Been Shown To Be Bioequivalent or Equivalent in Safety and Effectiveness to the Innovator Product for Which It May Be Substituted

1. A Difference in Bioavailability Between Premo's Allopurinol Product and the Innovator Product Could Have Serious Health Consequences

190. Certain cancer chemotherapy causes a rush of uric acid into the kidneys after treatment, which has a potential for crystallizing in the kidneys and significantly impairing renal function (Tr. at 1633-34 (Dr. Rhodes)). (A)

191. If an allopurinol product is less bioavailable than expected, there could be a build up of uric acid in the kidneys and resultant urinary failure, particularly in patients treated with allopurinol in connection with chemotherapy (Tr. at 1635 (Dr. Rhodes)); Tr. at 47-48, 45 (Dr. Crout)).

192. Some literature suggests that allopurinol may be useful in removing some of the cancer chemotherapy drugs which are themselves potentially carcinogenic (Tr. at 1634-35 (Dr. Rhodes)). (A)

2. Allopurinol Presents a Bioavailability Problem

193. Allopurinol presents a bioavailability problem. (Tr. at 605 (Dr. Cabana); see also Tr. at 843 (Dr. Shangraw)).

194. Dr. Rhodes suggested that a bioavailability study be performed with respect to allopurinol (Tr. at 1637 (Dr. Rhodes)). (A)

195. Dr. Rhodes was asked to design in a general way a bioavailability study for Premo's allopurinol product (Tr. at 1432-35; cf. 1602 (Dr. Rhodes)). (A)

196. There is no appropriate dissolution standard for allopurinol to assure the bioavailability of allopurinol tablets without an in vivo bioavailability test (P. Ex. 2 (Aff. of Dr. Cabana), ¶ 43.a). (NDBI)

3. Comparison of the Formulation of Premo's Allopurinol with That of the Innovator Product Provides no Basis for Concluding That Those Products Would Not Differ in Bioavailability

197. Plaintiff has shown it is not determinable whether the ingredients in Premo's allopurinol products differ in any respects from those in the innovator products (P. Ex. 12; P. Ex. 14; Tr. at 816-17 (Dr. Shangraw)).

Paragraph 198 is contained in the sealed portion of these findings.

199. Without knowing the intrinsic dissolution of the allopurinol drug substance, or its wettability, it is not predictable that the Premo product will be equivalent to the innovator product it imitates (Tr. at 836, 837 (Dr. Shangraw)). Wettability is the ability of water or gastric fluid to come into contact with a molecule of the drug on the surface of the solid particles; if the surface will not wet, then the particle will not dissolve (Tr. at 839-40).

200. Dr. Rhodes' conclusion that Premo's allopurinol product is therapeutically equivalent to the innovator product is based in part on data produced by United States Testing Co. that the products contain the same inert components present in the same or very similar concentrations, (Tr. at 1406 (Dr. Rhodes)), a conclusion not supported by the actual United States Testing Results ( see sealed findings 127-28, supra).

E. Premo's Doxine (Doxylamine Succinate Plus Pyridoxine Hydrochloride) Product Differs in Composition from the Innovator Product and Has Not Been Shown To Be Bioequivalent or Equivalent in Safety and Effectiveness to the Product It Copies

1. A Lack of Bioequivalence Between the Premo Product and the Innovator Could Have Significant Health Effects

201. This drug combination is used in pregnant women to combat nausea and vomiting ("morning sickness"). (A)

An effective overdosage because of bioavailability differences would result in adverse reactions such as dizziness, headache, and drowsiness while underdosage would result in therapeutic failure (P. Ex. 6 (Aff. of Dr. Perel), ¶ 15). (NDBI)

202. The question whether the product doxylamine succinate with pyridoxine hydrochloride is a drug having high or low therapeutic ratio is the subject of some debate (Tr. at 1360-61 (Dr. Rhodes)). (NDBI)

203. The fact that the active ingredients of a combination drug are separately safe and effective — as the active ingredients of this combination are alleged by Premo to be — would not indicate that the combination of those ingredients will be safe and effective. The drugs may interfere with each other or there may not be a proper medical rationale for the combination that would justify the increased risk from the two products (Tr. at 696-97 (Dr. Cabana)).

2. This Product Has Potential Bioavailability Problems

204. Premo's product containing doxylamine succinate with pryidoxine hydrochloride imitates an innovator product which is a controlled release dosage form, designed to delay the release of the drug substance over 8 to 10 hours. Controlled release dosage forms prevent severe bioavailability problems (Tr. at 577-78 (Dr. Cabana); P. Ex. 2 (Aff. of Dr. Cabana) ¶ 44.e). (NDBI)

3. Differences Between the Premo Formulation and That of the Innovator Raise Questions Concerning the Bioequivalence of These Two Products

205. The ingredients in Premo's doxylamine succinate plus pyridoxine hydrochloride product differ, in some respects, from those in the innovator product (P. Ex. 12; P. Ex. 17). (NDBI)

Paragraphs 206 through 207 are contained in the sealed portion of these findings.

208. Even if the active ingredients in this drug in a simple formulation would not provide significant problems, a formulation involving a slow release rate, such as the one found here, raises questions in terms of equivalence (Tr. at 829 (Dr. Shangraw)).

4. It Can Not Be Concluded That Premo's Doxine Product Is Equivalent to the Innovator Product

209. Because there has been no comparison of in vivo and in vitro data for Premo's doxylamine succinate with pyridoxine hydrochloride product, no firm conclusion concerning the bioequivalence of Premo's product to the innovator product can be made (Tr. at 1384 (Dr. Rhodes)).

F. Premo's Hydroxyzine Pamoate Product Differs in Composition from the Innovator Product and Has Not Been Shown To Be Bioequivalent or Equivalent in Safety and Effectiveness to the Product It Copies

1. A Lack of Bioequivalence Between Premo's Hydroxyzine Pamoate Product and the Innovator Could Have Significant Health Effects

210. Hydroxyzine pamoate is used in the management of anxiety, tension, and psychomotor agitation, during episodes of emotional stress and occasionally for the control of nausea. It is commonly used as an adjunct to psychotherapy. (A)

The failure of this drug to work because its bioavailability is too low, or because unexpected adverse reactions result from too high bioavailability, can result in setting back the effects of therapy that would, in some cases, otherwise allow afflicted patients to lead a reasonably normal life (P. Ex. 6 (Aff. of Dr. Perel), ¶¶ 12, 16).

211. A significant overdosage of this active ingredient because of bioavailability problems could cause drowsiness, tremors and, rarely, convulsive episodes (P. Ex. 6 (Aff. of Dr. Perel), ¶ 13). (NDBI)

212. An unintentional underdosage of hydroxyzine because of bioavailability problems would result in a therapeutic failure of the drug in the patients in question. Patients treated with the drug would thus be subject to anxiety, in some cases severe anxiety, which might interfere with the progress of their psychotherapy (P. Ex. 6 (Aff. of Dr. Perel), ¶ 14). (NDBI)

2. This Drug Presents a Bioavailability Problem

213. Hydroxyzine pamoate is an extremely insoluble salt and poses a good potential for a bioavailability problem (Tr. at 606 (Dr. Cabana); Tr. at 827 (Dr. Shangraw)). (NDBI)

214. In addition, because the hydroxyzine pamoate is labeled as equivalent to the hydroxyzine hydrochloride, it is necessary to show that the product is bioequivalent to hydroxyzine hydrochloride (Tr. at 606-07 (Dr. Cabana)).

215. Dr. Rhodes is aware of no literature that says there is no documented bioequivalence problem with hydroxyzine pamoate (Tr. at 1220-21 (Dr. Rhodes)). (NDBI)

3. Differences Between the Premo Formulation and That of the Innovator Raise Questions Concerning the Bioequivalence of These Two Products

216. The ingredients in the Premo hydroxyzine pamoate product differ, in some respects, from those in the innovator product (P. Ex. 12; P. Ex. 32-A, B C).⁵

5 This finding is not disputed "as to inactive ingredients."

Paragraphs 217 through 218 are contained in the sealed portion of these findings.

G. Premo's Hydroxyzine Hydrochloride Product Has Not Been Shown To Be Bioequivalent or Equivalent in Safety and Effectiveness to the Product It Copies

1. A Lack of Bioequivalence Between Premo's Hydroxyzine Hydrochloride Product and the Innovator Could Have Significant Health Effects

219. Hydroxyzine hydrochloride is used in the management of anxiety, tension, and psychomotor agitation, during episodes of emotional stress and occasionally for the control of nausea. It is commonly used as an adjunct to psychotherapy. (A)

2. Differences Between the Premo Formulation and That of the Innovator Raise Questions Concerning the Bioequivalence of These Two Products

220. The ingredients in the Premo hydroxyzine hydrochloride product differ, in some respects, from those in the innovator product (P. Ex. 12; P. Ex. 16).⁶

6 This finding is not disputed "as to inactive ingredients."

Paragraph 221 is contained in the sealed portion of these findings.

222. In light of the differences in manufacture between the Premo and innovator hydroxyzine hydrochloride products, additional data would be necessary in order to show the bioavailability of the drug (Tr. at 824 (Dr. Shangraw)).

H. Premo's Betamethasone Valerate Product Has Not Been Shown To Be Bioequivalent or Equivalent in Safety and Effectiveness to the Product It Copies

1. If This Product Is Not as Effective as Expected Because It Is Not Bioequivalent to the Approved Product That Fact Can Have Serious Consequences

223. If Premo's betamethasone valerate cream is not effective, resort to a more potent oral drug would be a real possibility (Tr. at 1260 (Dr. Kanof)). (NDBI)

2. This Drug Has Potential Bioequivalence Problems

224. Betamethasone valerate cream is a drug with a potential bioequivalence problem. There is relatively little absorption of drugs from ointments through the skin and there is some evidence indicating significant effects of the formulation on the absorption of the drug at the site of application. A highly lipid ointment applied against an aqueous membrane may cause the drug to stay in the ointment rather than getting into the site of drug action on or through the skin (Tr. at 610 (Dr. Cabana)).

225. Dr. Rhodes agreed that a clinical study of betamethasone valerate should be done (Tr. at 1436). (A)

3. Comparison of the Formulations of the Premo Product and That of the Innovator Raise Questions Concerning the Bioequivalence of Those Products

226. The ingredients in Premo's betamethasone valerate cream differ, in some respects, from those of the innovator product (P. Ex. 12; P. Ex. 18).

Paragraph 227 is contained in the sealed portion of these findings.

228. Lack of knowledge concerning the particle size of the active ingredient of the betamethasone valerate products and the degree of occlusiveness of the products make it impossible to determine that Premo's product is equivalent to the innovator product (Tr. at 830-31).

XII. There Is No Dispute That a Significant Number of Drugs Have Bioequivalence Problems and That the FDA's Standards for Concluding That Bioequivalence Testing Is Needed Are Reasonable

229. Dr. Rhodes' estimate of the number of prescription drugs with bioequivalence problems was limited only to "less than say 25%" (Tr. at 1264). (A)

230. Dr. Rhodes disputes that it is possible to determine therapeutic equivalence in all products by in vitro tests (Tr. at 1615 (Dr. Rhodes)), and there are many occasions when it is essential to conduct a biological availability study (Tr. at 1616). (A)

XIII. What Testing Premo Has Performed of Its Products Fails To Show Those Products Equivalent to the Innovators for Which They May Be Substituting

A. Premo's Clinical Study (D. Ex. 53) of Its Triamthiazide Study Does Not Show That Product To Be Bioequivalent or Equivalent in Safety and Effectiveness to the Innovator Product

231. The clinical trial submitted by Premo as comparing Premo's Triamthiazide with Dyazide in fifty subjects has the following flaws —

a. The trial does not evaluate the therapeutic effectiveness of the triamterene component of the Premo combination with respect to its intended use ( see P. Ex. 80-A, page 1; P. Ex. 80-B at 115 (Deposition of F. Ballantyne)) in preventing the development of low blood potassium or treating low blood potassium if it occurs, in that there was no measurement of serum potassium in the study (Tr. at 119-20 (Dr. Crout); Tr. at 155 (Dr. Cortell); see also Tr. at 1331 (Dr. Rhodes)).

b. The electrocardiograms said to have been done at two points in this trial would not be capable of measuring serum potassium at levels that would, if they were higher or lower than normal, be significant (Tr. at 127-29 (Dr. Crout); Tr. at 155-56, 172-73 (Dr. Cortell)). The acid-base status of a patient would affect the ability of electrocardiograms to reveal serum potassium levels (Tr. at 173 (Dr. Cortell)). In fact, by the time potassium level effects appeared on electrocardiograms, there would be real safety problems for the subjects of the test (Tr. at 128 (Dr. Crout); Tr. at 169-71 (Dr. Cortell)).

c. The test was not double-blind, that is the investigator was not blinded and because the investigator knew which patients were getting which drugs at which times there was a potential for subtle observer bias (Tr. at 121 (Dr. Crout)). The test was not even single-blind (P. Ex. 80-A, page 4; P. Ex. 80-B at 198 (Deposition of F. Ballantyne)).

d. A sequencing effect would occur, that is any effect going on simultaneously with the trial could influence the results from one drug more than the results from another. For example, a flu epidemic or a seasonal hot spell that occurred during a part of the trial in which the results of one drug was being measured might result in an incorrect evaluation of the effects of the drugs (Tr. at 121, 123 (Dr. Crout)).

e. The trial is not very sensitive or discriminating in that it does not use a placebo group and it is not possible to know whether the patients used as subjects had their blood pressure under control (Tr. at 123 (Dr. Crout)).

f. The investigators violated their own entrance criteria by using fourteen subjects whose blood pressure varied more than 20 mm systolic over 10 mm diastolic, i.e., patients whose blood pressure was not sufficiently controlled to satisfy the entrance requirements of the test (Tr. at 124 (Dr. Crout)).

g. The report of the test does not say whether the subjects were using drugs which varied in their dosage during the trial, which variance would be a fatal flaw in the trial (Tr. at 124-25 (Dr. Crout)).

h. The trial was inadequate even to determine whether the Premo product was equivalent to the approved product in its effect as an antihypertensive drug because the report does not reveal the variability of blood pressure in each patient, the circumstances under which the blood pressures reported were measured, or a determination that the blood pressures of the subjects had been adequately controlled on the first medication before the subjects were switched to the second medication (Tr. at 154 (Dr. Cortell)).

i. To the extent that this study shows anything, it shows a difference in the effectiveness of the Premo product as opposed to the approved product in controlling hypertension, with the Premo product appearing to be more effective in decreasing blood pressure (Tr. at 156 (Dr. Cortell); see also D. Ex. 53, page 6). This increase in potency of the Premo product, if real, could lead to hypotension (decreased blood pressure (Tr. at 159)), a condition that would be a cause for concern to the treating physician in patients who might be switched from the approved product to this product (Tr. at 156-59 (Dr. Cortell)).

232. The methods used in the clinical trial performed by Dr. Ballantyne do not meet the rigorous standards that would be needed for a full clinical trial of a new drug substance (Tr. at 1318-19 (Dr. Rhodes)).

233. Dr. Ballantyne, who performed the study on Premo's Triamthiazide product, admits that that study did not establish the influence of that product, or of the reference product, Dyazide, on serum potassium levels (P. Ex. 80-A, page 4; P. Ex. 80-B at 199 (Deposition of Dr. Ballantyne)).

234. Even Premo's own witnesses finally declined to support the position that the failure of EKG's taken in this study to show increases in potassium with patients treated with Triamthiazide showed that Premo's product did not significantly increase potassium levels.

235. Hyperkalemia, i.e., by Dr. Bloom's definition serum potassium of greater than 5.5 milligrams per deciliter (Tr. at 434), would probably not result in any changes on electrocardiograms (Tr. at 433 (Dr. Bloom)). (A)

236. An EKG "might" show changes in potassium level in blood serum if the figure was 7 milliequivalents of potassium per liter of blood (Ballantyne deposition at 170). A patient with a figure of 6 milliequivalents of potassium per liter of blood is in a serious state ( id.). Patients have died with a potassium figure of less than 6 (Ballantyne deposition at 171). A potassium reading of 6 requires serious medical attention and would cause Dr. Ballantyne to be quite alarmed if he saw that figure in his patients in a routine physical examination ( id.). A patient with a figure of a 6 milliequivalents of potassium per liter of blood would probably be in the hospital already (Dr. Ballantyne deposition at 172). Thus, before an EKG would show changes in potassium levels, the patient should already be in the hospital ( id.). (NDBI)

237. Dr. Rhodes is unqualified to disagree with the statements referred to, supra, made by Dr. Ballantyne in his deposition (Tr. at 1662 (Dr. Rhodes)). (NDBI)

238. There may be no electrocardiographic changes with slight elevations in serum potassium concentration and, at best, an imperfect correlation with moderate or large elevations — electrocardiographic changes usually appear only when the serum potassium exceeds 6.5 to 7.4 milliequivalents per liter (P. Ex. 230, page 1656), a fact that Dr. Rhodes does not question (Tr. at 1663-64 (Dr. Rhodes)). (NDBI)

239. Another source relied upon by Dr. Rhodes (D. Ex. 141, page 397) states that until recently there have been no reports of electrocardiographic changes occurring in patients who develop hyperkalemia from triamterene or the combination of triamterene and hydrochlorothiazide (Tr. at 1665 (Dr. Rhodes)). (A)

240. Premo's witnesses Dr. Rhodes and Dr. Tessler would say only that the Premo clinical study of Triamthiazide performed by Dr. Ballantyne showed therapeutic equivalence in reducing blood pressure (Tr. at 1320 (Dr. Rhodes); Tessler Deposition at 107).

241. Premo's Dr. Tessler agrees that an electrolyte measurement in the clinical study performed for Premo on its Triamthiazide product would have been useful in the context of showing that there was no material change from switching from the innovator product to Triamthiazide (P. Ex. 86-A, page 2; P. Ex. 86-B at 158-59 (Deposition of Dr. Tessler)).

242. Because the reference standard on the EKG charts used in this study was, in some instances, cut off, any physician other than Dr. Ballantyne could not verify whether those charts were consistent with hyperkalemia and thus would simply have to take his word for it (P. Ex. 80-A, page 4; P. Ex. 80-B at 178-79 (Deposition of Dr. Ballantyne)).

243. The failure to measure potassium levels in the blood, the most important parameter in determining whether use of Premo's Triamthiazide as a substitute for Dyazide is safe, was the result of a conscious decision to perform this study where potassium levels could not be measured:

a. If it were easy to obtain a measurement of serum potassium with respect to the 50 patients in the clinical tests of Premo's Triamthiazide drug, that would be the prudent thing to do (P. Ex. 80-A, page 2; P. Ex. 80-B at 135).

b. Dr. Ballantyne and his technicians are accomplished in spinning down blood into serum for transport to the United States for analysis (P. Ex. 80-B at 116). After being spun down the serum is frozen and then transported to the United States (P. Ex. 80-A, page 1; P. Ex. 80-B at 116-117). When serum is sent to the United States in a frozen state it is possible to test for electrolytes, such as potassium levels (Ballantyne deposition at 117). The test is a simple one (P. Ex. 80-A, page 1; P. Ex. 80-B at 117).

c. Blood serum that was, according to the protocol, to be transported to the United States for analysis was not transported because Biometric Testing, Inc. did not provide an individual to carry it from the Virgin Islands to the United States (P. Ex. 80-A, page 2; P. Ex. 80-B at 118-20).

d. Dr. Ballantyne was aware that the laboratory that would analyze serum samples from his test, Rockland Medilabs, had the ability to perform electrolyte testing on those serum samples at the time that he was involved in consulting on the design and carrying out of a clinical test on Premo's Triamthiazide drug (P. Ex. 80-A, page 2; P. Ex. 80-B at 127).

e. Electrolyte analysis of blood taken in St. Vincent's at a laboratory in the United States is possible if the blood has been spun down into serum. However, an impediment to such analysis is the danger that some of the blood would hemolyze during transport between clinics in the countryside, which do not have a centrifuge to spin down blood into serum, and the Kingston General Hospital on St. Vincent's, which does have a centrifuge (P. Ex. 80-A, page 2; P. Ex. 80-B at 136).

f. The only real piece of equipment you need to spin down blood into serum is a centrifuge, which is not a very expensive piece of medical equipment (P. Ex. 80-A, page 2; P. Ex. 80-B at 135). There is only one centrifuge on St. Vincent's Island (id.). Biometrics could have provided centrifuges to Dr. Ballantyne for use in outlying clinics if it chose to do so (Tessler deposition at 162-63).

g. If Dr. Ballantyne had had centrifuges in his clinics in the countryside he could have spun the blood down and preserved the serum until it could be frozen and shipped to the United States (P. Ex. 80-A, page 2; P. Ex. 80-B at 128). If he had brought the patients into the hospital in Kingston he could also have done so (P. Ex. 80-A, page 2; P. Ex. 80-B at 128-29).

h. About half of the patients in the study were drawn from the Kingston General Hospital and, with respect to those patients, the blood could have been spun down into serum without any danger of hemolyzing and thus could have been analyzed for electrolytes (P. Ex. 80-A, page 2; P. Ex. 80-B at 133-34). This was not, however, done ( id.).

i. Dr. Tessler knew that a study done in the Virgin Islands that was not centralized in Kingston would not have facilities to do the testing for potassium levels in the blood (P. Ex. 80-A, page 3; P. Ex. 80-B at 151-53; P. Ex. 86-A, page 2; P. Ex. 86-B at 159-60, 161-62 (Deposition of Dr. Tessler)). Yet, Dr. Tessler did not discuss with Dr. Ballantyne centralizing this test in Kingston nor did he discuss the electrolytes problem at all (P. Ex. 80-A, page 3; P. Ex. 80-B at 153-54 (Deposition of Dr. Ballantyne)).

244. Dr. Tessler does not know that the Triamthiazide used in the clinical study of Triamthiazide performed by Biometric Testing, Inc., was from a production batch of Triamthiazide (P. Ex. 86-A, page 1; P. Ex. 86-B at 143 (Deposition of Dr. Tessler)) and the record contains no proof that it was.

245. There is no clinical study showing the effectiveness of Premo's triamterene with hydrochlorothiazide product in edemous patients (Tr. at 1666 (Dr. Rhodes)).

B. The "Potassium Sparing" Study (P. Ex. 22-A) of Premo's Triamthiazide Does Not Demonstrate That Premo's Product Does Not Cause Greater Potassium Loss than the Innovator Product

246. The study submitted by Premo that purports to evaluate the short term effect of Premo's combination product containing triamterene and hydrochlorothiazide on serum potassium and on urinary excretion of potassium (P. Ex. 22-A) is insufficient to demonstrate that the triamterene component of Premo's Triamthiazide would fulfill its function of preventing potassium build up (Tr. at 131 (Dr. Crout)). This study (P. Ex. 22-A) has the following flaws:

a. This study does not address the question of the effects of chronic administration of Premo's Triamthiazide on potassium build up in the body, which is the question of importance in evaluating the effectiveness of the triamterene component of Triamthiazide (Tr. at 131, 177 (Dr. Crout); Tr. at 160-61 (Dr. Cortell)).

b. The urinary excretion of potassium was reported only in milliequivalents per liter of urine, a measurement that does not tell how much actual potassium was excreted in the urine because the volume of urine is not reported (Tr. at 139-41, 177-79 (Dr. Crout)).

c. The results reported are, at least in some instances, internally inconsistent. For example, in Figure One of P. Ex. 22-A, the serum potassium levels are falling while there is no increase in urinary potassium, a finding that is internally inconsistent and makes no sense (Tr. at 141-42 (Dr. Crout)).

d. One of the thirteen patients in the study was dropped from the study without any explanation of the criteria for selecting the patient that was dropped (Tr. at 142-44 (Dr. Crout); P. Ex. 84-A; P. Ex. 84-B at 175, 343 (Dep. of N. Robillard)).

247. If this test was intended to measure very precisely differences in potassium-sparing activity of Premo's product as opposed to the innovator product, it would have been useful to have a longer test and to have a test with more subjects (Tr. at 1333 (Dr. Rhodes)).

248. In the potassium-sparing study of Premo's Triamthiazide product, the patients could have been taking other medication, i.e., estrogen therapy as a means of contraception, and that fact was not reported (P. Ex. 84-A, page 2; P. Ex. 84-B at 341-42 (Deposition of N. Robillard)). (NDBI)

249. This study was not carefully done. The investigators failed to take obvious precautions to protect the subjects. Female subjects of childbearing age were used, an unusual practice that worries concerned physicians (Tr. at 160-61 (Dr. Cortell)). The tests for pregnancy used in the female subjects in this study would not pick up pregnancy early in the first trimester, even though most teratological damage, i.e., damage to the fetus, occurs in the first trimester of pregnancy (P. Ex. 84-A, Page 2; P. Ex. 84-B at 342 (Deposition of N. Robillard)). It has been reported that hydrochlorothiazide is a known penetrator of the placental barrier (P. Ex. 84-A, page 2; P. Ex. 84-B at 340 (Deposition of N. Robillard)). In this study of Premo's Triamthiazide product, one of the patients had a potassium level of 6.1 milliequivalents per liter, a level that by Mr. Robillard's definition indicates hyperkalemia (P. Ex. 84-A, page 1 2; P. Ex. 84-B at 200, 354 (Deposition of N. Robillard); P. Ex. 86-A, page 2; P. Ex. 86-B at 183 (Deposition of Dr. Tessler)), and that Dr. Ballantyne believes would require hospitalization (Ballantyne Deposition at 172). At the time when this potassium level was observed, Clinical Concepts was administering that patient a double dose of triamterene plus hydrochlorothiazide (P. Ex. 84-A, page 3; P. Ex. 84-B at 355 (Deposition of N. Robillard)). Dr. Tessler has never prescribed in his practice a single dose of Dyazide (amounting to 100 milligrams of hydrochlorothiazide in 200 milligrams of triamterene) which would be four capsules in a single administration (P. Ex. 86-A, page 2; P. Ex. 86-B at 175-76 (Deposition of Dr. Tessler)). Giving multiple doses of drugs to normal subjects, in and of itself raises ethical questions (Tr. at 160 (Dr. Cortell)).

C. The Incomplete Bioavailability Study of Premo's Triamthiazide Does Not Demonstrate That Premo's Product Is Bioequivalent or Equivalent in Safety and Effectiveness to the Innovator Product

250. Dr. Rhodes was willing to provide an outline of a protocol for a bioavailability study of triamterene plus hydrochlorothiazide but was not asked to do so (Tr. at 1602 (Dr. Rhodes)). (NDBI)

251. The triamterene and hydrochlorothiazide plasma levels reported from a bioavailability study performed by Premo (D. Ex. 83-A and 84) do not reflect completed studies as the study was set out in its protocol (Tr. at 1667-68 (Dr. Rhodes); P.Ex. 82-A and B, 114-15 (Deposition of Dr. Cohn); and Tr. at 2185).

1. The Study Can Not Be Relied upon Because It Was Improperly Done⁷

7 This Government heading read: "The Study Can Not Be Relied Upon Because It Was So Sloppily Done".

252. The mean calculation in Table I is not done correctly (Tr. at 941-42 (Dr. Barr)). The errors made would tend to show that Premo's product was more equivalent to the innovator than if the calculations had been done correctly (Tr. at 942). The standard deviation and the coefficient of variation and some of the other statistical terms are also miscalculated (Tr. at 942-43).

253. There were questionable interpolations of some of the most critical data points in the original report of this study when the actual data did not exist (Tr. at 943-44 (Dr. Barr)).

254. Dr. Greene felt it would be "difficult to give an honest and accurate estimate" of two points missing for a subject in the triamterene plasma level section of the Triamthiazide bioavailability study (D. Ex. 83), yet those points were estimated in the original report of that study.

255. The triamterene plasma level section of the bioavailability study of Triamthiazide, even as revised by Dr. Greene (P. Ex. 58; D. Ex. 83 Revised)), is a study whose overall validity is thrown into question by its questionable implementation (Tr. at 972, 973, 975 (Dr. Barr)).

2. The Reports of Triamterene Plasma Levels, Even as Revised by Dr. Greene, Do Not Show Equivalence of That Component

256. Neither the original Premo triamterene plasma level report from this bioavailability study (P. Ex. 46) nor the revision of that report (P. Ex. 58 (also identified as Ex. 83 Revised)) would prove the equivalence of the two products tested (Tr. at 946 (Dr. Barr)).

a. The Study Was Simply Too Insensitive To Detect a Difference in Triamterene Plasma Levels if One Exists

257. By definition a power analysis calculates the sensitivity of a bioavailability test (Tr. at 1906-07 (Dr. Greene); see also Tr. at 1672 (Dr. Rhodes)). (A)

258. The Premo bioavailability study of its Triamthiazide product has a 12% power to detect a 20% difference between the Premo product and the innovator product in triamterene plasma levels (Tr. at 1985-86 (Dr. Greene); P. Ex. 601).

259. A power of 15% for the triamterene blood plasma levels would be extremely low (Tr. at 1674 (Dr. Rhodes)).

260. Dr. Rhodes did no power analysis or any other type of analysis of the data in the triamterene plasma report of the bioavailability study of Premo's triamterene with hydrochlorothiazide product, rather he "eye-balled" it and based his conclusion upon that process together with looking at the average values obtained (Tr. at 1685-86 (Dr. Rhodes)).

261. It appears from the data in this study that Premo's product is more bioavailable than the innovator product, though one cannot draw a conclusion on this subject because no power analysis was performed (Tr. at 944-45 (Dr. Barr)).

262. The power of this study to detect significant differences between the two drugs was very low (Tr. at 945 (Dr. Barr)).

263. A revised analysis of a test lacking sufficient power to detect a difference cannot add to the power of that test (Tr. at 946-47 (Dr. Barr)). (NDBI)

264. In the triamterene plasma study, as revised (P. Ex. 58), there is tremendous variation in subjects — subject 9 has a blood level of 156 while subject 17 has a blood level of 7 — that does not lead to any finding of difference in the test, indicating that the test is not particularly sensitive to differences (Tr. at 976-77 (Dr. Barr)).

265. The Premo triamterene plasma study has such great variability that a test with sufficient power to determine whether there are real differences is extremely important (Tr. at 979 (Dr. Barr)).

266. There is less variability in the Premo triamterene plasma study with respect to the innovator product than with respect to the Premo product, but the test is so insensitive that conclusions can not be fairly drawn from that fact (Tr. at 980 (Dr. Barr)).

267. Any investigator beginning a bioavailability study for a product containing triamterene should have known that there was considerable intra and inter subject variation with respect to bioavailability for that drug (Tr. at 1598-99 (Dr. Rhodes); cf. Tr. at 1897 (Dr. Greene)), a fact that would have a tendency to lower the sensitivity of the test (Tr. at 1600 (Dr. Rhodes)), yet there is no evidence of attempts to make this test more sensitive to compensate for this predictable problem.

b. Problems with the Assay Used May Have Accounted in Part for the Lack of Sensitivity of this Bioavailability Test

268. Fluorometric assays for bioavailability studies are very difficult to do and provide many sources of error (Tr. at 963 (Dr. Barr)). (A)

269. Both Dr. Rhodes and Dr. Greene suggested that an HPLC technique be used as an analytical method for triamterene in the triamterene plus hydrochlorothiazide bioavailability study rather than the method actually used, a fluorometric technique (Tr. at 1653-54 (Dr. Rhodes); P. Ex. 245, Tr. at 1655-56 (Dr. Rhodes)). (NDBI)

270. The HPLC assay has as one of its characteristics high specificity, and that is one of the reasons why. Dr. Greene preferred it (Tr. at 1906 (Dr. Greene)). (NDBI)

271. At the time when Dr. Greene recommended the HPLC assay rather than the fluorometric assay, he knew that the fluorometric assay was being considered and had reviewed an article that explained that the fluorometric assay differentiated triamterene from its metabolites (Tr. at 1902-06 (Dr. Greene)). (NDBI)

3. The Reports of the Hydrochlorothiazide Levels in this Study, even as Revised by Dr. Greene, Do Not Show Premo's Product To Be Equivalent to the Innovator

272. When asked whether Premo's bioavailability studies (D. Ex. 84) demonstrate that Premo's product is bioequivalent to the innovator product with respect to the hydrochlorothiazide component of the two products, Premo's witness Dr. Rhodes did not state that the study showed the two products to be equivalent. Rather he stated that the data reported "does not indicate any statistical difference in the biological availability for the hydrochlorothiazide from the Smith Kline Product and from the Premo product" (Tr. at 1310 (Dr. Rhodes)). This is the language contained in FDA's definition of "bioequivalent drug products." This definition, contained at 21 C.F.R. § 320.1(e) states that bioequivalent products means pharmaceutical equivalence or pharmaceutical alternatives whose rate and extent of absorption do not show a significant difference when absorbed at the same molar dose of the therapeutic moiety under similar, experimental conditions, either single-dose or multiple-dose.

a. This Study Was Simply Too Insensitive To Detect a Difference in Hydrochlorothiazide Plasma Levels if One Existed

273. The study that compared hydrochlorothiazide plasma levels produced by Premo's Triamthiazide product with those produced by the innovator product (P. Ex. 59 (also D. Ex. 84)) did not have the power to detect a real difference between the two products (Tr. at 947-49 (Dr. Barr)). With respect to the reported relative bioavailability of the two products, the plasma levels from Premo's product are reported to be greater than those from the innovator product in some subjects by as much as two-fold with an area under the curve in which the Premo product is 147% of the area under the curve of the innovator product. While because the test is not sufficiently sensitive to conclude that the difference is a real difference (Tr. at 950 (Dr. Barr)), the variability of the study warrants that it be discarded. As Dr. Barr testified, it has less than 50% chance of detecting differences as much as 50% ( id.). Dr. Greene's testimony to the contrary is rejected.

274. The power of the Premo bioavailability study of its Triamthiazide product to detect a 20% difference in hydrochlorothiazide levels is 13% (Tr. at 1993 (Dr. Greene); P. Ex. 602)), a very low power (Tr. at 1994 (Dr. Greene)). By definition this means that in only 13 out of 100 cases will this test detect differences of 20% between the mean value for the test drug and the mean value for the reference drug ( id.).

b. Questions About the Incomplete Report of the Assay Used To Measure Hydrochlorothiazide Made the Reliability of the Reported Results Questionable

Paragraphs 275 through 277 are contained in the sealed portion of these findings.

278. At least with respect to two examples in Dr. Cohn's preliminary report of the hydrochlorothiazide assay, a sample was reported as containing less hydrochlorothiazide than a sample known by Dr. Cohn to contain no hydrochlorothiazide (Tr. at 2178 (Dr. Cohn); D. Ex. 403, subject # 9, subject # 11). (NDBI)

279. Dr. Cohn relied upon results from which were stricken certain numbers which he believed were outliers (numbers which were more than two standard deviations from the mean), although the numbers stricken were not in fact more than two standard deviations from the mean (Tr. at 2188-90 (Dr. Cohn); D. Ex. 405-B). The only number properly stricken as an outlier was one number in the 50 nanogram range (Tr. at 2190-91, 2192). A re-analysis by Dr. Cohn, including the improperly stricken values, showed a difference with respect to some values (Tr. at 2276-77 (Dr. Cohn); D. Ex. 405-B, 405-BB). While this re-analysis established that the differences were extremely small, and, according to Dr. Cohn, would not affect the conclusions reached in his analysis, the fact remains that the improper assay standard values were used in the analysis of the data.

4. There Must Be a Valid in Vivo Bioavailability Study of Triamthiazide To Show Its Bioequivalence to the Innovator

280. In vitro dissolution tests cannot form the basis for conclusions that Premo's Triamthiazide product is equivalent in bioavailability to the marketed product, Dyazide, because there is no statistical correlation between in vitro dissolution and in vivo bioavailability for the triamterene component and there is no other information that establishes that the dissolution tests are predictive of in vivo bioavailability of this type of drug. (A)

D. The Bioavailability Study of Premo's Chlorthalidone Product Does Not Show Premo's Product To Be Bioequivalent or Equivalent in Safety and Effectiveness to the Innovator Product

281. The Premo chlorthalidone bioavailability study performed by Biometric Testing Inc., to determine that the Premo chlorthalidone product is bioequivalent to Hygroton, the approved chlorthalidone product (Tr. at 538 (Dr. Cabana)), is rejected as unacceptable.

282. The Premo chlorthalidone bioavailability study (P. Ex. 60 (also Defendants' Ex. 39-A)), through a revised analysis, has a number of flaws. Thus, there are a large number of missing data points in the study and the statistical analysis indicates that the power of the study to detect a difference if one exists was relatively small and too low to validate this study (Tr. at 951-53 (Dr. Barr)). The defendants' suggestion that Dr. Greene compensated for missing data points and corrected arithmetic error to the satisfaction of the government's own witness, Dr. Barr, is rejected.

1. The Chlorthalidone Study Has So Many Inaccuracies That It Cannot Be Considered Valid

283. As Dr. Cabana stated, Premo's chlorthalidone study had many subjects with missing values (Tr. at 644).

The United States has offered as a Proposed Finding that the Biometric Testing conviction concerned false statements and writings supplied to the Food and Drug Administration by a totally separate and distinct division of Biometric Testing, Inc. None of the personnel involved in the Premo chlorthalidone study were convicted. The two divisions were physically separate ( See Robillard Dep. at 232-236 (Ex. D-68-B); Tessler Dep. at 118-121, 124).⁸

8 For the reasons set forth, the conviction, although admitted, has been given no weight.

284. The chlorthalidone bioavailability study performed by Biometrics for Premo contained transposition errors.

285. The physical gathering of data is extremely important in a bioavailability test, particularly with respect to blood samples taken during the absorptive phase of the drug and at or near the maximum concentration (the point at which the drug would be expected to reach its peak level in the blood), because such values cannot be estimated properly if the samples are not correctly taken (Tr. at 502-05 (Dr. Cabana)). The Premo chlorthalidone study (P.Ex. 21) was flawed because of transposition errors between the raw data and the data used in the computer. Transposition errors in the urine data, some of them significant, appeared in five different subjects. Transposition errors, again some of them significant, appeared in reports of the study of blood on five occasions, involving three patients (Tr. at 508-10 (Dr. Cabana)).

286. The Premo chlorthalidone study, performed by Biometric Testing, Inc., has many missing values in urine, both in terms of drug content and volume of urine estimated (Tr. at 513-23 (Dr. Cabana)).

287. The Premo chlorthalidone study involved exclusion of subjects from the trial without an explanation of the criteria for determining which subjects should be excluded or why. Without defined criteria, deletion of patients introduces the possibility of bias into the study (Tr. at 534-35 (Dr. Cabana)).

288. Estimation of a point on the down slope of the curve in urine in the chlorthalidone study is totally invalid (Tr. at 647-49 (Dr. Cabana)). (NDBI)

289. In the chlorthalidone study the amount estimated for urine concentration is higher than would be expected because a relatively high amount of urine would normally lead to a lower concentration of drug in the urine (Tr. at 650 (Dr. Cabana)). (NDBI)

290. The Premo chlorthalidone study measured three parameters, the C-max, the T-max, and the area under the curve (AUC). While the summary data indicated the means are similar, evaluation of the mean results of a bioavailability study can be very misleading because there may be a large difference in variability about the mean (Tr. at 956 (Dr. Barr)). Thus biopharmaceuticists would not make a comparison based on these data alone. In the Premo chlorthalidone study there was a pattern of variability with respect to both the Premo and the innovator products (Tr. at 957 (Dr. Barr)).

291. In a conversation with Walter Epler, on or about January 10, 1980, Dr. Rhodes expressed concern about certain samples in the chlorthalidone bioavailability study performed by Biometric Testing, Inc. (Tr. at 1630-31 (Dr. Rhodes)). (NDBI)

292. Dr. Rhodes in a letter dated January 14, 1980, referring to Premo's chlorthalidone bioavailability study stated, "We are concerned that in a number of subjects a significant number of samples were either not taken or were lost (for example see subject # 30). We don't understand the reasons for this." (P. Ex. 243, page 2; Tr. at 1690 (Dr. Rhodes)). Dr. Rhodes stated his opinion in the January 14, 1980, letter that the Premo bioavailability study for chlorthalidone would not be acceptable to the Food and Drug Administration because "the December 20, 1979, guidelines [on chlorthalidone] have not been complied with." What Dr. Rhodes was referring to here is the fact that Premo initiated its chlorthalidone bioavailability study using Hygroton 100 mg as a reference product. This was appropriate under the guidelines in effect when the study began. Several months later when FDA decided to change its reference product to oral solution, the Premo study became unacceptable (P. Ex. 243; Tr. at 1691 (Dr. Rhodes)).

293. Dr. Rhodes suggested that another bioavailability study on Premo's chlorthalidone product be begun in January of 1980, yet he was unaware whether any such study had been undertaken at the time of his testimony in June 1980 (P. Ex. 243, page 2; Tr. at 1692 (Dr. Rhodes)). (A)

294. Dr. Rhodes agrees that when he stated in his sworn affidavit that "Premo has commissioned a well-designed and executed bioavailability study in which their product was compared with the bioavailability of the innovator's product," he was wrong (Tr. at 1694 (Dr. Rhodes)). (A)

2. Dr. Greene's Reanalysis of This Study Was Based on Questionable Interpolations

295. While it is difficult to fit the curve where data points are missing on the ascending portion of that dose response curve (Tr. at 1955 (Dr. Greene)) (A), Dr. Greene did so in approximating the values of missing data points in the chlorthalidone study (Tr. at 1952-53). As Dr. Barr suggested, the techniques used by Dr. Greene rendered his results unreliable (Tr. at 944 (Dr. Barr)).

296. Dr. Greene interpolated a value on subject # 3 of D. Ex. 39-A that represented two important parameters, T-Max and C-Max of the results for that subject (Tr. at 1956-57), i.e., he interpolated two of the three main parameters in reviewing the results of this test (Tr. at 1957 (Dr. Greene)). He did the same thing for subjects # 13 and # 15, and subject # 25 (Tr. at 1957-59). With respect to subject # 25, the figure that Dr. Greene filled in jumped the data point from 9.36 to 10.24 in a two-hour period (Tr. at 1959). Between the six- and eight-hour period interpolation has the effect of moving the drug level from 7.95 to 10.24, a jump of 2.29 (Tr. at 1959-60 (Dr. Greene)). (NDBI)

3. The Power of the Chlorthalidone Study, even Assuming It Were Valid, Would Be Insufficient To Detect a Difference Between Premo's Product and the Innovator

297. Even using the values reported by Biometric, the "beta error" is such that there is less than a 50% chance of detecting a difference between the Premo products and the reference product if such a difference exists (Tr. at 523-24 (Dr. Cabana)). Similarly, if one analyzes the test by comparing the results in each patient who receives the various products, there is less than a 47% probability of seeing any difference that may exist ( i.e., power) using that perspective as well, a totally unacceptable probability, far beyond what statisticians would ever accept (Tr. at 524-27 (Dr. Cabana)).

298. Less than one-third of the subjects in the Premo chlorthalidone study (P. Ex. 21) fell within acceptable boundaries, i.e., above 75% of the reference and below 125% of the reference (Tr. at. 529-30 (Dr. Cabana)). Notwithstanding defendants' attack upon Dr. Cabana, I found him to be a well-qualified expert and credit his testimony.

299. The data in the Premo chlorthalidone study, which Dr. Cabana found "confounding," if accepted at face value, leads to the conclusion, probably not a correct one, that two of Premo's 50 milligram chlorthalidone products are not equivalent in bioavailability to one of Premo's 100 milligram products (Tr. at 532-33, 523-24 (Dr. Cabana)).

4. A Different Reference Standard Should Have Been Used

300. The FDA requires a bioavailability test of chlorthalidone to compare the tested drug against an oral solution of chlorthalidone (Tr. at 487 (Dr. Cabana); P. Ex. 4, (Aff. of Dr. Crout), Ex. 6). Dr. Rhodes agrees that it would be more useful to compare Premo's chlorthalidone product to an aqueous solution or suspension of the drug substance than to compare it to the marketed product (Tr. at 1342 (Dr. Rhodes)). An oral solution is used because there may be variation in the approved product that would make comparison to that product undesirable (Tr. at 487-88 (Dr. Cabana)). Thus, Premo's chlorthalidone bioavailability study (P. Ex. 21), fails to meet an important, though not crucial, design criterion in that it does not compare Premo's product to a standard of known performance (Tr. at 497 (Dr. Cabana)). (NDBI)

301. If Premo had run the chlorthalidone study using a solution as the reference standard, rather than using the approved chlorthalidone product, the "noise" or "variance about the blood levels in the individual subjects with time" would have been reduced (Tr. at 641-42 (Dr. Cabana)).

E. The Study of Premo's Betamethasone Valerate Cream Is Incomplete and of Flawed Design

302. Dr. Kanof's study is not completed (Tr. at 1262) and, as far as he is aware, there has been no statistical analysis on the data gathered up to this point (Tr. at 1261). (NDBI)

303. The phenomenon of "observer bias" is a tendency which may sometimes occur for an observer in a clinical trial to find what he is looking for or is expected to find. (A) This is ordinarily an unconscious reaction by the observer and is a well-known phenomenon among clinical pharmacologists. (Tr. at 1248 (Dr. Kanof)). In the study in process, Dr. Kanof, the observer, knew that the thesis he was testing would be proved if there were no difference between the two products used on each patient (Tr. at 1257-58 (Dr. Kanof)). If a placebo had been used in the test, that would have introduced some uncertainty with respect to that proposition (Tr. at 1257), but there was no placebo in Dr. Kanof's study (Tr. at 1255).

F. The Pilot Study of Premo's Allopurinol Product Is Insufficient To Show the Equivalence of Premo's Product to the Innovator

304. The Premo pilot study dealing with its allopurinol product does not provide sufficient data to draw a definitive conclusion as to the bioequivalence of the Premo product to the innovator product (Tr. at 1290 (Dr. Rhodes)). (NDBI)

XIV. Defendants' Contentions Concerning Injury to Themselves if an Injunction Is Issued Are Exaggerated; Any Losses Suffered Are Attributable to a Business Decision To Market Drugs That Defendants Knew the Government Believed To Be Illegal

305. Premo has approved new drug-applications or certifications for the marketing of antibiotics, and for 22 other drug products, and Federal Pharmacal has such applications approved by the Food and Drug Administration for 22 more products (P. Ex. 5 (Affidavit of Dr. Crout), paragraph 3).⁹

9 Since the time of trial, FDA has also approved an ANDA for a Premo chloropromide product, a formulation different from that declared not to be a "new drug" by the Court in Premo Pharmaceutical Labs., Inc. v. United States, 475 F. Supp. 52 (S.D.N.Y. 1979), rev'd, 629 F.2d 795 (2d Cir. 1980).

306. The United States notes in its Proposed Findings that Premo keeps no business records separated for particular drugs; and thus there is no documentary support for Mr. Silverang's testimony that Premo purchased various equipment that can be used only for drugs that the government believes should be the subject of an injunction (Tr. at 2080 (Silverang)). I need not resolve this issue. If there is such equipment, Premo purchased it at its own risk and the injury it will thereby sustain as a result of an injunction does not outweigh the factors favoring an injunction.

307. According to Mr. Silverang the five tablets named as examples in the complaint in this action represent 12 million of the 44 million tablets Premo's sales forecast projects will be sold each month (P. Ex. 85-A, page 2; P. Ex. 85-B at 93). The hydroxyzine pamoate and Triamthiazide capsules represent 12 million of the 40 million capsules the sales forecast projects will be sold per month ( id.). The betamethasone valerate cream represents 3,100 kilograms of the approximately 6,100 kilograms a month of ointments and creams Premo projects marketing (P. Ex. 85-A, page 2; P. Ex. 85-B at 94.)¹⁰

10 As the United States notes in its Proposed Findings, there is apparently an error in the transcript since Mr. Silverang computes the 3,100 kilograms as about 21% of the ointments and cream production (P. Ex. 85-A, page 2; P. Ex. 85-B at 94).

308. The Parke-Davis encapsulating machines Premo uses are rented (P. Ex. 85-A, page 2; P. Ex. 85-B at 151-52). If Premo decided to remove these machines after failing to renew the yearly lease, Premo would not have to pay for their use (P. Ex. 85-A, page 2; P. Ex. 85-B at 152). Premo could have leased the Osaka encapsulating machine through a sale-lease-back arrangement but chose not to do so (P. Ex. 85-A, page 2; P. Ex. 85-B at 154-55). It decided to purchase this machine in part because of the substantial tax savings from an investment tax credit associated with buying rather than leasing the equipment (P. Ex. 85-A, page 2; P. Ex. 85-B at 155-56). Premo could have rented more Parke-Davis encapsulating machines rather than buying the Osaka machine (P. Ex. 85-A, page 2; P. Ex. 85-B at 156-57). (NDBI)

309. In January of 1980 Premo spent approximately $65,000 for change parts for an encapsulating machine that is capable of making one million capsules a day (Tr. at 2074 (Silverang)). These change parts would allow for making capsule sizes different from those for which the machine had previously been used ( id.). These change parts would not be used for Triamthiazide, since that product had already been manufactured by the machine (Tr. at 2075). While Mr. Silverang stated that the change parts could be used for the product hydroxyzine pamoate (Tr. at 2076), Premo already has five encapsulating machines with a capacity of 750,000 capsules per day capable of handling the production of 100,000 capsules per day of hydroxyzine pamoate Premo manufactures (Tr. at 2055-56 (Silverang)). (NDBI)

310. Premo ordered a larger Hycoater the day after this law suit was commenced, March 13, 1980 at a cost of $110,000 (Tr. at 2091-92 (Silverang)). (NDBI)

Mr. Silverang is aware that the Government has seized a number of the products named as examples in the Government's complaint and that FDA has contested Premo's assertion that Premo can market these drugs free of regulatory action (Tr. at 2096 (Silverang)). (NDBI)

311. In January of 1978 Premo executives discussed the marketing of betamethasone valerate cream and decided that the FDA requires preclearance for that product (P. Ex. 85-A, page 2; P. Ex. 85-B at 176-77 (Deposition of S. Silverang)). In that meeting the executives also discussed the fact that trifluoperazine requires FDA pre-clearance before marketing (P. Ex. 85-A, page 2; P. Ex. 85-B at 178). In deciding whether to market a drug product, it is standard operating procedure for Premo executives to consider whether that product needs to be precleared by the Food and Drug Administration (P. Ex. 85-A, page 2-3; P. Ex. 85-B at 179, 190 (Deposition of S. Silverang). (NDBI)

312. In deciding whether or not to market the drug allopurinol Premo executives knew that FDA required preclearance for that drug but did not discuss, so far as Mr. Silverang knows, whether studies would be performed in man or in animals to test the safety and effectiveness of the Premo allopurinol product (P. Ex. 85-A, page 3; P. Ex. 85-B at 183-84 (Deposition of S. Silverang)). (NDBI)

313. The eight drugs named as examples in the Government's complaint counted for only 5% of total sales for the period November 1978 through February 1979 (P. Ex. 26-A, page 1; P. Ex. 26-B at 100-02 (Deposition of John Blackman); P. Ex. 26-C-3). Only three of the eight drugs were being marketed at that time ( id.). (NDBI)

314. The statement that 55% of the total gross sales of Premo's products were attributable to the eight drugs named as examples in the Government's complaint is based in large part upon a million dollars worth of sales of Premo's Triamthiazide product in November and December of 1979 (P. Ex. 26-A, page 2; P. Ex. 26-B at 104-105). (A)

315. The findings above, 306-312, demonstrate that Premo's injury from an injunction could have been avoided had it sought FDA advice before making the capital investments to which reference is made.

XV. Defendants' Contentions That They Are "Responsible Drug Manufacturers" Are Belied by the Record

316. Clinical Concepts had submitted a protocol to Premo for a bioavailability study of Premo's Triamthiazide product on or about August 29, 1979 (P. Ex. 84-A, page 1; P. Ex. 84-B at 248 (Deposition of N. Robillard)). In an August 30, 1979, letter describing the protocol for these pilot studies, Mr. Robillard stated, inter alia, that: "Each of these pilot studies will have the added advantage of providing Premo with preliminary information of the bioequivalence of their formulation of the respective drugs relative to the trade-name reference products" (P. Ex. 84-A, page 1; P. Ex. 84-B at 253-54; P. Ex. 39-B).

317. A pilot study performed by Clinical Concepts for Premo Pharmaceuticals in two runs compared Dyazide to a Premo Triamthiazide product from Lot 11754, apparently an experimental lot (P. Ex. 39-D; P. Ex. 84-B at 274-77). Lot 11754 produced an area under the curve reported as .23 while the Dyazide value was .095 (P. Ex. 39-D). These results were provided to Walter Epler of Premo on or about September 1979 (P. Ex. 84-B at 276-77).

318. Either before or after that pilot study was completed Clinical Concepts did a comparison in which four subjects were tested. Two received a solution, one received a capsule apparently from an experimental lot of Triamthiazide, lot 11760, and the fourth subject, # 2, was administered a drug from Premo lot # 099014 (P. Ex. 84-B at 275; P. Ex. 39-D; P. Ex. 26-C-5). The fact that lot # 099014 of Premo's Triamthiazide was used in this test is confirmed by the P. Ex. 39-H in which "Drug A" is identified as triamterene plus hydrochlorothiazide capsules, lot # 099014 (P. Ex. 84-A, page 2; P. Ex. 84-B at 283-84 (Deposition of N. Robillard); see also P. Ex. 81-A, page 1; P. Ex. 81-B at 22-23 (Deposition of J. Blackman); P. Ex. 61 (Aff. of Robillard), page 5). The area under the curve obtained for subject # 2 was 6.933 (P. Ex. 39-D). The figures from this test were reported to Premo Pharmaceuticals by Mr. Robillard (P. Ex. 84-A, page 2; P. Ex. 84-B at 278 (Deposition of N. Robillard)). The results of this study appear correctly in P. Ex. 26-C-5 (P. Ex. 84-A, page 2; P. Ex. 84-B at 279-80 (Deposition of N. Robillard)), received by Premo before the beginning of 1980 (P. Ex. 81-A, page 1; P. Ex. 81-B at 23 (Deposition of J. Blackman)).

319. There can be no dispute that lot # 099014 was a production batch of Triamthiazide marketed by Premo. This batch was so identified by defendants to the Court in their examination of Dr. Rhodes (Tr. at 1292 (Dr. Rhodes); see D. Ex. 113).

320. The only product in the pilot bioavailability study that is identifiable as being seven to eight times more bioavailable than Dyazide is product 099014, the marketed drug (P. Ex. 26-C-G5; P. Ex. 39-D).

321. Mr. Epler submitted an affidavit in federal district court in SK F Co. v. Premo Pharmaceutical Laboratories, Inc., 481 F. Supp. 1184 (D.N.J. 1979), in which he stated that there have been preliminary testing of Premo's Triamthiazide product and that:

On the Basis of These Preliminary Test Results, Mr. Robillard of Clinical Concepts Stated That Premo's Product, "Triamthiazide, Has a Higher Absolute Bioavailability than SKF Co.'s Product, Dyazide, and That It May Well Be Significantly More Bioavailable Which Can Be Confirmed upon Complete Bioavailability Testing" (P. Ex. 4 (Aff. of Dr. Crout), Ex. 5).

322. Mr. Epler stated that the conclusion he expressed in that affidavit was based upon a statement by Mr. Robillard that Premo's product may well be significantly more bioavailable than the Dyazide product upon completion of bioavailability testing (P. Ex. 27-A, page 3; P. Ex. 27-B at 26 (Deposition of W. Epler)). Mr. Epler stated that Mr. Robillard had told him that one patient was given a capsule from a production batch of Triamthiazide and that the numerical value (indicating level of drug in the blood) for that capsule was higher than the numerical value for the Dyazide capsule (P. Ex. 27-A, pages 3-4; P. Ex. 27-B at 26-27).

323. After the FDA questioned Premo about the results of this test, which had been reported in a protocol submitted to the agency, Premo took the public position articulated in Mr. Seymour Blackman's affidavit, see quotation supra (P. Ex. 67, Affidavit of S. Blackman, ¶ 9, page 5).

324. Mr. Blackman, on the stand in this case, stated that he had sworn to the affidavit submitted in the Southern District of New York without checking to see whether the statement he had made was true (Tr. at 2248 (S. Blackman)). Rather, Mr. Blackman testified that he asked Mr. Epler a question that appears to have been calculated not to elicit an answer that would inform him whether the Premo production batch was the one that showed the high bioavailability. He asked Epler if Premo had sold a batch "wherein we designed a sevenfold availability into our product" (emphasis added) and Mr. Epler said, "Of course not" ( id.).

325. Walter Epler states that he reported to Seymour Blackman the fact that Triamthiazide had a higher absolute bioavailability than SKF's Dyazide (P. Ex. 27-A, page 4; P. Ex. 27-B at 27-28 (Deposition of Epler)).

326. Premo did a potassium sparing study of its Triamthiazide product only because of "bad press" from FDA concerning the potential hazards associated with that product (P. Ex. 25-A, page 5; P. Ex. 25-B at 61 (Deposition of J. Blackman)).

XVI. Evidence in the Record Impeaches Various Premo Witnesses Suggesting That Less Weight Should Be Accorded to Their Testimony

327. Mr. Silverang stated that the decision to market trifluoperazine was made in October of 1979 (P. Ex. 85-B at 29 (Deposition of S. Silverang)), yet Premo had manufactured two very large batches of trifluoperazine much earlier. One, of a million tablets, should have cleared the tableting machine 2 or 3 working days after March 28, 1979 (P. Ex. 81-A, page 1; P. Ex. 81-B at 3 (Deposition of J. Blackman)). The tableting of another, with a theoretical yield of 3 million tablets, began on February 7, 1979 (D. Ex. 99; P. Ex. 81-A, page 1; P. Ex. 81-B at 4-5 (Deposition of J. Blackman)). (NDBI)

328. Dr. Greene, in preparing his supplemental analysis of the chlorthalidone bioavailability study, testified that he prepared the study when he "noted that there were transposition errors in the entry of the raw data points into the computer" (Tr. at 1746 (Dr. Greene); see D. Ex. 39 and 39-A). He prepared this study, however, only after those transposition errors had been pointed out by government witnesses at the trial which he attended. Prior to those trial days he had endorsed the chlorthalidone study (D. Ex. 77 (Aff. of Dr. Greene), ¶ 6). Similarly, Dr. Greene stated that he eliminated certain patients from the study because "I felt that it was difficult to estimate what the concentrations in blood would be if there were what I considered in my professional opinion to be an excessive number of missing points within a single run" (Tr. at 1751 (Dr. Greene)). Although he stated in the affidavit filed in this court on March 19, 1980, that he had examined the interim bioavailability data from this study, he did not point out these deficiencies in his professional opinion at that time (D. Ex. 77; ¶ 7). (NDBI)

329. Dr. Greene testified on direct examination that the power of the triamterene blood level part of the Triamthiazide bioavailability test to detect a difference as small as a .1% was in the neighborhood of 30 to 50% (Tr. at 1847), yet he admitted on cross-examination that the power of this test to detect a 20% difference was 12% (Tr. at 1985-86), clearly a logical inconsistency. Similarly, with respect to the hydrochlorothiazide results, Dr. Greene stated that the power of the test to see whether there was a "very, very, very small difference, insignificant difference, was of 30 to 50%" (Tr. at 1863-64), yet he said the power of this test to find the differences of 20% was 13% (Tr. at 1993).

330. Though Dr. Greene believed that drugs with dissolution rates below 1 milligram per milliliter per cubic centimeter have serious absorption and bioavailability problems, he did not, in deciding whether triamterene, hydrochlorothiazide, or chlorthalidone had bioavailability problems, consider whether those drugs fit within those parameters (Tr. at 1900-01 (Dr. Greene)). (NDBI)

331. In Biometric's analysis of its bioavailability study of Premo's chlorthalidone product, the analysis of variance included subject variation (Tr. at 1997 (Dr. Greene); D. Ex. 39). The Clinical Concepts computer analysis of the triamterene blood levels and hydrochlorothiazide in its bioavailability study of Premo's Triamthiazide includes subject variation in the analysis of variance (Tr. at 2031-33 (Dr. Greene); D. Ex. 83, ¶ 601). Professor Westlake, upon whom Dr. Greene relied, states that subject variation is one of the potential sources of variation that must be accounted for before proper tests on the formulation effects of drug effects can be made (Tr. at 2000-02 (Dr. Greene); P. Ex. 250, pages 159-60). Dr. Wagner, upon whom Dr. Greene relied, states that one of the sources of variability is intersubject variability and another is intrasubject variability (Tr. at 2004, 2007 (Dr. Greene); P. Ex. 181, page 290). Dr. Greene, in the analysis of variance he prepared in his reanalysis of the chlorthalidone and triamterene studies, did not account for intersubject variability (Tr. at 2032 (Dr. Greene); D. Ex. 83; Revised; 84; 39-A).

332. Dr. Greene, in computing the power of the chlorthalidone study as found in Defendant's Exhibit 600, did not follow the Wagner formula correctly (Tr. at 2046-47 (Dr. Greene)).

333. Dr. Rhodes, in relying upon defendants' exhibit 8, drew conclusions other than those drawn by the author, in reliance upon an article that was missing the page containing the author's conclusions, on the basis of charts referring to other sources which Dr. Rhodes did not check (Tr. at 1642-46 (Dr. Rhodes; D. Ex. 8; P. Ex. 241).

334. Douglas S. Greene, Ph.D., suggested in his Affidavit of July 7, 1980, that inclusion of intersubject variability in the residual or error term of an analysis of variance was supported by two texts he recognized as authoritative. Dr. Greene agreed, however, that the Johnson Leone text he had cited suggested that the expected value of each of the residual of random variables was zero and that based upon his review of the initial Biometric Chlorthalidone study he was aware that intersubject variation reported as a separate assignable term was statistically significant. Thus its expected value was not zero. (A)

Dr. Greene also agreed in the Deposition that there was a typographical error in the second article which he relied upon "Statistical Inferences Applied to Bioavailability: A Guide for the Practicing Pharmacist" which significantly changed the meaning of the passage quoted in his Affidavit. Dr. Greene further admitted that he did not note this error when preparing his Affidavit. (P. Ex. 703a; 703b at 9-35 (Deposition of Dr. Greene); D. Ex. 610 at pp. 3-62; P. Ex. 701).

335. Dr. Greene stated in his Affidavit of July 7, 1980, that he performed a power analysis with respect to the Chlorthalidone data reported in Defendant's Exhibit 39a and, using three separate analyses involving tests for power, found in each case, the power of the test to detect a 20% difference to be in excess of 99%. However, in his first case analysis reported in Paragraph 12 of his Affidavit Dr. Greene postulated that three means representing the products — Hygroton 100 mg., Premo 50 mg. Chlorthalidone X2, and Premo 100 mg. Chlorthalidone — each were dispersed 20% from the overall mean. During his Deposition Dr. Greene admitted that it was not physically possible to have three means expressed around an overall mean with the same 20% difference from the mean. With respect to Dr. Greene's power analysis reported in Paragraph 13 of his Affidavit, Dr. Greene stated that the procedure for performing a power analysis suggested by Dr. Bernard Cabana made use of the mean of the reference product, i.e., the Hygroton 100 mg. tablet, rather than the overall mean in performing the analysis as endorsed by Dr. Wagner. Dr. Greene based this conclusion on Dr. Cabana's article "Bioavailability of Imipramine Tablets Relative to a Stable Isotop-Labeled Internal Standard: Increasing the Power of Bioavailability Tests" (Appendix B to D. Ex. 610). Dr. Greene during his Deposition, however, agreed that Dr. Cabana's application of the power analysis formula made use of the overall mean concept as endorsed by Dr. Wagner rather than the mean of the reference product so suggested by Dr. Greene. Dr. Greene admitted, therefore, that his statement in Paragraph 13 of his Affidavit was incorrect. Dr. Greene postulated at Paragraph 14 of his Affidavit a case three analysis to evaluate the power of a bioavailability test. That analysis hypothesized a product A being 80% of the reference mean and product B being 120% of that reference. Dr. Greene denied that this case involved a computation of the power of a test to compute a 40% rather than a 20% difference in the particular means. He agreed, however, that the purpose for an analysis of variance is to confirm or reject the null hypothesis, i.e. the hypothesis that all three means in this example are equal, thus mean A equals mean B equals mean R. If the null hypothesis is satisfied drug A will equal drug B. Dr. Greene agreed that the F test does in effect compare all three means to determine whether the null hypothesis is satisfied. (P. Ex. 703a; 703b at 37-53 (Deposition of Dr. Greene); D. Ex. 610 pp. 12-14).

336. Paragraph 15 of the Affidavit of Douglas S. Greene, Ph.D., was stricken pursuant to agreement by counsel for the parties.

337. Defendants presented nine expert witnesses at the hearing. These witnesses are:

a) Christopher T. Rhodes, Ph.D. Dr. Rhodes is professor and chairman of the Department of Pharmacy at the University of Rhode Island. He is a pharmaceutical scientist who specializes in the formulation and processing of pharmaceutical dosage forms, quality control, and evaluation of pharmaceutical products. His special areas of expertise additionally encompass the evaluation of inactive ingredients to determine if they are suitable for use in pharmaceutical products and the formulation of drugs into compressed tablets. He has published approximately 100 research papers in the field, and is co-editor of two books entitled: Automated Methods for the Analysis of Drugs and Other Substances of Pharmaceutical Interest; and Modern Pharmaceutics — Dosage Form, Design and Evaluation. T-1030-1036; Ex. D-76. Dr. Rhodes has extensive experience in the design and evaluation of bioavailability studies of pharmaceutical products.

b) Douglas S. Greene, Ph.D., is an assistant professor of pharmacy at the University of Rhode Island. As a clinical pharmacokineticist, Dr. Greene is experienced in the use of mathematics to determine the processes that occur after a drug is administered. Dr. Greene teaches graduate and undergraduate courses in pharmacokinetics and biopharmaceutics. He is currently conducting research as a member of Brown University (Roger Williams General Hospital Clinical Cancer Center) where his research is supported by two grants from the National Cancer Institute. He is also currently funded by several federal grants to develop new assays for drugs. Dr. Greene analyzes bioavailability data for pharmacists, pharmaceutical companies, and physicians. He has numerous publications and presentations in the area of biopharmaceutics and pharmacokinetics T-1710-1716; T-1879-1880; Ex. D-77.

c) Norman B. Kanof, M.D., a physician specializing in the practice of dermatology, treats patients on a daily basis suffering from dermatologic disorders. Dr. Kanof graduated from George Washington University Medical School in 1941. He trained for three years with the New York Skin Cancer Unit from 1945 to 1948, and a year of fellowship training which lead to a Doctorate of Medical Sciences in Dermatology. He has been in private practice in New York since 1949. He is an associate professor of dermatology at the New York University School of Medicine and an attending dermatologist at New York University Hospital and at St. Clair's Hospital, New York City. He is familiar with drug products used to treat conditions that occur in the practice of dermatology, and is familiar with the conduct and design of clinical trials designed to show the safety and effectiveness of drug products. Over the past ten years, he has conducted approximately 75 to 100 studies as an investigating physician, the great majority of which have been with topical products. T-1231-1234, 1246; Ex. D-79.

d) Jerry Bloom, M.D., received a degree in medicine from Down State University, University of New York, and trained as an intern and resident in internal medicine Maimonides Hospital in Brooklyn, New York, and Boston City Hospital. He received advanced training in infectious diseases and internal medicine at Boston City Hospital and the Ohio State University Hospital of Ohio State University. He is Board certified in internal medicine. He has teaching appointments at the Down State Medical Center as an associate professor of clinical medicine and an appointment at NYU School of Medicine as clinical assistant professor of medicine. He is familiar with the treatment of hypertension, and his practice deals with the treatment of patients suffering from hypertension. T-373-381; Ex. D-78.

e) Max A. Tessler, M.D., graduated from New York University, Bellevue Medical School in 1955, interned in medicine, Fourth Medical Division, Bellevue House, from 1955 to 1956, and served as assistant resident in internal medicine, Bellevue Hospital, 1956 to 1957. Dr. Tessler was the chief resident in medicine at N Y University Hospital in 1959, and from 1959 to 1960 he was a trainee and research fellow in gastroenterology at Yale Medical Center. Dr. Tessler has been in private practice since 1962. He has been certified by the National Board of Medical Examiners, and the Specialty Board of Internal Medicine. Dr. Tessler is on the teaching staff at N Y University Hospital and also at the Manhattan Veterans' Administration Hospital. He is a consultant at Cabrini Medical Center. The last two years of Dr. Tessler's residency programs and fellowships involved conducting clinical studies both at Yale Medical Center and subsequently at the Cornell Division of Bellevue. During this two-year period, Dr. Tessler acquired considerable experience working with outstanding researchers in terms of the design of studies, the conduct of clinical studies, the management of patients during clinical studies, and the collection of data during clinical studies. Dr. Tessler has been conducting clinical studies himself for over twelve to thirteen years. He has been involved in Phases I, II, III, and IV research. Tessler Dep. at 1-11 (Ex. D-69B).

f) Frederick N. Ballantyne, M.D., is the chief of staff, medical staff, in the General Hospital in St. Vincent, and senior medical officer for the Island Health Services, the Island of St. Vincent. As chief physician, he is in charge of the medical staff and also in charge of the medical or internal medicine wards at the hospital. Dr. Ballantyne's practice is in internal medicine, and he is involved in the regional health clinic program on the island. Dr. Ballantyne graduated from Howard University in 1959 magna cum laude, and received his medical training in Upstate Medical Center in Syracuse from 1959 to 1963. He is Board Certified in internal medicine. Dr. Ballantyne is experienced in clinical trials and research, and in the conduct of clinical trials for drugs used to treat hypertension.

g) Mr. Sidney Kahan is the president of Fittelson Laboratories, Inc., New York, New York, and has been associated with Fittelson Laboratories for 20 years. He holds a bachelor of science, Columbia University, 1935, a chemical engineering degree, Columbia University, 1936, and a master of science in organic chemistry, Polytechnic Institute of New York, 1949. Mr. Kahan worked for the U.S. Food and Drug Administration from 1938 to 1955 where he was primarily engaged in laboratory analysis of food and drugs. He also inspected food plants for compliance with FDA regulations, sanitary quality, and good manufacturing practices. Mr. Kahan is an expert in analytical chemistry for the purposes of performing quantitative and qualitative analyses on components of drug products.

h) Richard Cohn, Ph.D., received his bachelor of science in pharmacy from Temple University's School of Pharmacy in 1966. He has a master of science degree in toxicology that was awarded to him by Temple University Graduate School in 1968. He has a doctor of philosophy in pharmacology obtained from Jefferson Medical College, Thomas Jefferson University, in 1974. He is a licensed pharmacist in both the states of Pennsylvania and New Jersey. Dr. Cohn is presently associated with National Medical Services, Inc. as a laboratory director. National Medical Services is a private professional toxicology service. Dr. Cohn has had eight years of experience as a duty toxicologist with the Poison Information Center of the City of Philadelphia in disseminating information concerning exposure to potentially toxic substances, both to lay individuals and to clinicians, both private and within hospital environments. For the past five years, he has been intimately involved with the standardization and review of bioavailability data with eventual submission of reports to FDA. Dr. Cohn's education, training, and experience are in the fields of toxicology, pharmacology, and bioanalytical assay work. He has conducted tens of thousands of bioanalytical assays, and has developed and validated dozens of bioanalytical assays. Cohn Dep. at 1-18 (Ex. D-66B).

i) Norman F. Robillard, M.T., is president of Clinical Concepts, Inc., Englewood Cliffs, New Jersey. Clinical Concepts is a corporation which contracts with the food, cosmetic, and pharmaceutical industries to conduct clinical studies for the purposes of evaluating food, cosmetic, or drug products. Mr. Robillard is a medical technologist and thus is trained in the theory and practical application of conducting analysis on bodily fluids for the purposes of establishing disease processes. The medical technologist's position is a certified position which requires appropriate training and internship, and a national examination administered by the American Medical Technologists' Association. Mr. Robillard has been involved in the conduct of bioavailability studies for approximately nine to ten years. Within that time, he has developed somewhere in the neighborhood of 400 to 500 protocols.

XVII. The Drugs in Issue

338. Premo produces and distributes in interstate commerce what are known as "generic" drugs, i.e., a drug product purporting to contain the same active ingredient as an established drug that has been marketed by the innovator company to a material extent and for a material time. Dr. Rhodes, T-1038; See Dr. Crout, T-213; Ex. D-75.

339. The eight drugs specifically named in the Complaint and alleged to be new drugs, and on which most of the testimony at the hearing centered, are as follows:

a) Betamethasone valerate .1% cream, marketed by Premo under the trade name "Beta Val," which is a generic version of the betamethasone valerate .1% cream product marketed by Schering Corporation under the trade name "Valisone." Ex. D-102, D-502, D-510; Ex. G-12-8, G-64.

b) Allopurinol tablets, in 100 and 300 mg dosage, marketed by Premo under the generic name, which is a generic version of the 100 and 300 mg allopurinol tablets marketed by Burroughs-Welcome Company under the trade name "Zyloprim." Ex. D-75, D-103, D-104, D-501, D-509; Ex. G-12-3, G-64.

c) Hydroxyzine hydrochloride tablets in 10, 25, and 50 mg strengths, marketed by Premo under the generic name, which is a generic version of hydroxyzine hydrochloride tablets marketed by Roering Division of Pfizer Pharmaceuticals under the trade name "Atarax." Ex. D-75, D-105, D-106, D-107, D-500, D-508; Ex. G-12-6, G-64.

d) Hydroxyzine pamoate capsules, 25, 50, and 100 mg, marketed by Premo under the trade name "Hy-Pam," which is a generic version of hydroxyzine pamoate 25, 50, and 100 mg capsules marketed by Pfizer under the trade name "Vistaril." Ex. D-75, D-108, D-109, D-109A, D-504, D-512; Ex. G-12-5, G-64.

e) Triamterene 50 mg with hydrochlorothiazide 25 mg capsules, marketed by Premo under the trade name "Triamthiazide," which is a generic version of triamterene 50 mg hydrochlorothiazide 25 mg capsules marketed by Smith, Kline and French under the trade name "Dyazide." Ex. D-75, D-111-130, D-506, D-514; Ex. G-12-1, G-64.

f) Chlorthalidone 50 and 100 mg tablets, marketed by Premo under the generic name, which is a generic version of the drug chlorthalidone 50 and 100 mg tablets marketed by USV Pharmaceutical Corp. under the trade name "Hygroton." Ex. D-75, D-100-101, D-505, D-513; Ex. G-12-2, G-64.

g) Doxylamine succinate 10 mg with pyridoxine hydrochloride 10 mg tablets marketed by Premo under the trade name "Doxine," which is a generic version of the doxylamine succinate 10 mg pyridoxine hydrochloride 10 mg tablet marketed by Merrell National Laboratories under the trade name "Bendectin." Ex. D-75, D-110, D-503, D-511; Ex. G-12-7, G-64.

h) Trifluoperazine hydrochloride 2 and 5 mg tablets, marketed by Premo under its generic name, which is a generic version of 2 and 5 mg trifluoperazine hydrochloride tablets marketed by Smith, Kline and French under the trade name "Stelazine." Ex. D-75, D-98, D-507, D-515; Ex. G-12-4, G-64. (A)

340. The active ingredients of the Premo Products ( i.e., those ingredients in a drug product which produce the desired pharmacological effect in the patient) purport to be the same quantitatively and qualitatively as the innovator products which they imitate. Dr. Crout, T-89; Dr. Rhodes, T-1202-1212; Ex. D-43, D-48, D-98-130; Ex. G-12-1-8, Ex. D-508-516, D-43, D-48; See, however, my Findings 28-30.

341. Seven of the Premo products bear the same labeling conditions for use as does the innovator product:

a) Hydroxyzine Hydrochloride Tablets. Ex. D-500 (Premo labeling); Ex. D-508 (Atarax labeling).

b) Allopurinol Tablets. Ex. D-501 (Premo labeling); Ex. D-509 (Zyloprim labeling).

c) Betamethasone Valerate Cream 0.1%. Ex. D-502 (Premo labeling); Ex. D-510 (Valisone labeling).

d) Doxylamine Succinate with Pyridoxine Hydrochloride Tablets. Ex. D-503 (Premo labeling); Ex. D-511 (Bendectin labeling).

e) Hydroxyzine Pamoate. Ex. D-504 (Premo labeling); Ex. D-512 (Vistaril labeling).

f) Chlorthalidone. Ex. D-505 (Premo labeling); Ex. D-513 (Hygroton labeling).

g) Triamterene and Hydrochlorothiazide. Ex. D-506 (Premo labeling); Ex. D-514 (Dyazide labeling).

h) With respect to trifluoperazine hydrochloride, the Premo product is labeled to be used only in the treatment of psychosis (Ex. D-507), whereas the Smith, Kline and French Stelazine brand of trifluoperazine is labeled for use in certain additional indications (Ex. D-515). The drug is not generally recognized as safe and effective for use in these additional indications, which the Food and Drug Administration has found the drug to be "less than effective." See 41 Fed.Reg. 37834, (Ex. G-4, Ex. 7), and Stipulation of the Parties, Ex. D-605. (A)

342. The innovator products of the eight drugs at issue are all the subject of approved NDA's. Dr. Crout, T-213; Ex. G-12-1-8. (A)

343. The defendants' contention (Dep. Proposed Finding) that each of the active ingredients of the eight drugs are generally recognized, among qualified experts, as safe and effective under the conditions of use prescribed, suggested, or set forth in its labeling, is rejected. The active ingredients of products cannot themselves be generally recognized as safe and effective, since those active ingredients are only used in drug products. Although some drug products, such as, for example, Premo's chlorthalidone, bear the name of the active ingredient, the products themselves are clearly different from the active ingredient alone. The testimony of Dr. Seife, relied upon by Premo, referred solely to recognition based on the performance of the marketed products, not to the active ingredients (P. Ex. 69-A, page 1; P. Ex. 69-B at 81 (Deposition of Dr. Seife)). The fact, testified to by Dr. Crout, that triamterene with hydrochlorothiazide is commonly used in patients is no evidence that that product is generally recognized as safe and effective since many approved new drugs are commonly used in patients.

344. A drug product is normally made up of the active ingredient (the drug) which produces the intended pharmacologic effect, as well as inactive ingredients, which are fillers, binders, lubricants, or other excipients. Dr. Shangraw, T-716-717. (A)

345. Each of the eight drugs at issue, both as produced by the innovator company and by Premo, contain inactive as well as active ingredients. Ex. D-98-130; Ex. G-12-1-8. (A)

346. I must reject the defendants' proposed finding 21 that, until recently, the Food and Drug Administration interpreted the new drug definition to mean that so long as the active ingredient of a drug had achieved general recognition of safety and effectiveness among qualified experts, any firm could market a drug product containing that active ingredient without the result that the second product became a "new drug." As argued by the United States, Mr. Byers, Associate Director for Compliance of the FDA's Bureau of Drugs, testified that firms inquiring into the new drug status of their products were told that certain of those drug products were not new drugs. Mr. Byers did not say that this determination was made upon a conclusion that the active ingredient of a drug had achieved general recognition of safety and effectiveness among qualified experts. Dr. Crout's testimony, as relied upon by Premo, dealt with a proposed policy, never implemented (Tr. at 70), that would have set up old drug monographs based upon a careful analysis of particular prescription drug products that would be similar to the FDA's action with respect to over-the-counter products ( see Tr. at 68-71). Dr. Crout did not say that the FDA had interpreted the new-drug definition to mean that so long as the active ingredient of a drug had achieved general recognition of safety and effectiveness among qualified experts, any firm could market a product containing that active ingredient without that drug becoming a new drug. The Hoffmann-LaRoche case referred to by Dr. Crout (Tr. at 69-70) dealt with an FDA enforcement policy that allowed unapproved new drugs to be marketed while abbreviated new-drug applications for those drugs were processed. That policy, enjoined by the Hoffmann-LaRoche Court, was based not upon a conclusion that the prescription products were not new drugs but rather upon what the agency believed at that time to be an appropriate enforcement policy. That policy was abandoned after the Hoffmann-LaRoche decision.

347. The Government contends that the term "new drug" applies to a drug product regardless of the fact that the active ingredient, or for that matter, inactive ingredients, have achieved general recognition of safety and effectiveness among qualified experts.

348. The inactive ingredients utilized by Premo in each of the eight drug products at issue are generally recognized among qualified experts as safe and effective and suitable for use in the compounding of finished pharmaceutical products. Dr. Cabana, T-634-635; Dr. Crout, T-254-282; Dr. Rhodes, T-1198, 1201, 1204, 1205. (NDBI)

349. The identity and quantity of the inactive ingredients utilized by Premo in production of the eight drug products at issue, and the quality control procedures and manufacturing practices and procedures utilized by Premo in the production of the eight drug products, are not generally known to the scientific community, as this information is considered proprietary and confidential. Dr. Rhodes, T-1418-1419. The same thing is true for virtually every finished pharmaceutical product produced by any firm in the United States. Dr. Rhodes, T-1418-19; Dr. Shangraw, T-867. (A)

350. The United States Pharmacopeia (USP) and the National Formulary (NF) are official compendia which prescribe standards of identity, strength, quality, and purity for some active ingredients, and in some circumstances, finished pharmaceutical products. Dr. Rhodes, T-1043; Dr. Shangraw, T-858; Ex. D-552-554. (A)

351. A drug is adulterated under § 501(b) of the Act, 21 U.S.C. § 351(b) (1976),

if it purports to be or is represented as a drug the name of which is recognized in an official compendium, and its strength differs from, or its quality or purity falls below the standards set forth in such compendium.

352. FDA's Good Manufacturing Practice regulations ("GMP's"), 21 C.F.R. Part 211, contain a code pertaining to the way in which drug products should be satisfactorily fabricated. These regulations provide a code regulating the production of drug products that, if taken together with pre-market clearance, assures that products are satisfactorily fabricated.

353. The fact that a generic drug contains the same active ingredient as an established drug does not, in and of itself, mean that the generic version will necessarily act in the same way in the treatment of disease as the established drug. Dr. Cabana, T-471. (A)

354. Under certain circumstances, it is possible that a generic version of a drug having the same active ingredient as an established drug may differ in bioavailability from the innovator drug. Dr. Cabana, T-473, T-655; see Dr. Crout, T-213-241. (A)

355. The following is a well established and accepted pharmaceutical principle, at least insofar as approved products are concerned:

If two or more drug products not presenting a bioequivalence problem contain identical amounts of the same active drug ingredient or therapeutic moiety in the same dosage form, are both manufactured in compliance with current good manufacturing practice, both contain inactive ingredients generally recognized as safe and suitable for the drug product formulation, both meet compendial or other standards of identity, strength, quality and purity and are both adequately labeled, it is reasonable to assume that these products will be of equivalent safety and effectiveness.

Ex. D-5, 42 Fed.Reg. 1625 (1977); Dr. Crout, T-218; Dr. Cabana, T-585; Dr. Rhodes, T-1050.

356. It is reasonable to assume equivalent safety and effectiveness based upon the factors set forth in Finding 355 regardless of the existence of an approved NDA or ANDA. Dr. Rhodes, T-1041; Dr. Crout, T-224. (A)

357. In vivo bioavailability data means data derived from a scientific study which measures the rate and extent of absorption of the active drug from a finished pharmaceutical product in man. Dr. Rhodes, T-1065; Dr. Greene, T-1822-1830; see Dr. Cabana, T-476; see Dr. Barr, T-882. (A)

358. Normally three parameters are measured in an in vivo bioavailability study:

a) T-max, which is the amount of time elapsed from initial administration of the drug until the time it reaches its greatest concentration in the blood;

b) C-max which is the amount of maximum concentration of the active drug in the blood; and

c) Area Under the Curve ("AUC"), which is the total amount of drug absorbed from the beginning to the conclusion of the test. Dr. Greene, T-1822, 27, 30; Dr. Rhodes, T-1965. (A)

359. In vitro bioavailability data means physical and chemical tests performed on a finished pharmaceutical product "in glass" or outside the body, and encompasses such testing as comparative dissolution studies (involving a measurement of the rate and extent of dissolution of the active drug substance from a pharmaceutical product into water or other media), disintegration testing, and the like. Dr. Rhodes, T-1053-1054; Dr. Barr, T-898. However, in vitro testing is often not predictive of bioavailability.

360. A clinical study of a drug involves administering the drug to human subjects and observing the clinical effectiveness and adverse reactions associated with the administration of the drug. See Dr. Barr, T-899. (A)

361. I reject the defendants' proposed finding 46 that "if two products contain the same drug substance and they are not bioequivalent, but the slope of the dose response curve is sufficiently flat, both products will produce the same therapeutic response and the same clinical effect." As the Government states, a drug may have different response curves with different slopes for its therapeutic response and for its various toxic responses (Tr. at 929-31 (Dr. Barr)), as is true with some of the drugs in issue here (Tr. at 935-36). Also, since no dose response curve that involves any response at all is flat over its entire length, the question of at what point the response is steep and at what point it is flat must also be addressed (Tr. at 915).

362. Therapeutic ratio means the ratio between the amount of the drug necessary to accomplish the intended therapeutic effect and the amount which will produce a toxic effect. Dr. Rhodes, T-1061. (A)

363. The higher the therapeutic ratio, the safer the drug is and the less concern there is about possible variability in blood levels between the two products. Dr. Rhodes, T-1061; Ex. D-5 at 1625. (A)

364. For example, the drug Digoxin, commonly used to treat patients with heart disease has a narrow therapeutic ratio, i.e., the difference between the amount of drug necessary to produce the desired therapy and the amount which will produce toxic effects is very small. Dr. Cabana, T-548; Dr. Bloom, T-391; Dr. Rhodes, T-1511-1512. (NDBI)

365. The Food and Drug Administration has established regulations requiring proof of bioequivalence which must be met by producers of all digoxin products. Dr. Cabana, T-551; Dr. Greene, T-1780. 21 C.F.R. § 310.500. (NDBI)

366. A drug does not present a medically significant bioequivalence problem solely because of a determination that a lack of bioequivalence could result in a serious adverse effect. Dr. Rhodes, T-1064-1065; Ex. D-5 at 1630. However, as Ex. D-5 notes, that factor would be "given great weight" in an FDA decision whether a bioequivalence requirement should be placed upon drugs that are approved by FDA and are legally on the market (42 Fed.Reg. at 1630, col. 2). The medical significance of a potential bioequivalence problem would suggest that concern about that problem is appropriate even though evidence of the potential bioequivalence problem is less than conclusive ( id.).

367. The standard operating procedures (SOP's) used by Premo in the production of each of the eight drugs in this case are quite detailed instructions developed by management so that the personnel would know exactly what to do in the production of a finished product. Dr. Rhodes, T-1196. (NDBI)^fn11

368. Premo's standard operating procedures are appropriate and compare very favorably with SOP's of other pharmaceutical companies. Dr. Rhodes, T-1196. (NDBI)

369. The United States agrees that the procedures and standards utilized for the production of the eight drugs at issue would be perfectly acceptable if they had passed through premarket approval. Dr. Rhodes, T-1192-1196.

370. There is no dispute that Premo complies with good manufacturing practices. T-1192-1195.

a) Premo is required under GMP's to perform tests to make sure its products meet the USP. If the Premo product did not meet USP specifications, or if there were an absence of data that it met the USP, this would be called to the attention of the inspector's superior at FDA. Dr. Cabana, T-570-571.

b) There has been no allegation by the government that Premo products fail to comply with GMP's. Record, passim. (NDBI)

371. Premo's betamethasone valerate .1% cream is a generic version of Valisone brand of betamethasone valerate 1% cream produced by Schering Corporation. Dr. Crout, T-213, Ex. D-102, D-502, D-510; Ex. G-12-8. (A)

372. It is not disputed that these products' formulations show that they contain the same amount of an active ingredient labeled as betamethasone valerate. It is possible, however, for Premo's betamethasone valerate's active ingredient to differ from that used by the innovator.

373. Both products are labeled for use in the treatment of corticosteroid-responsive dermatoses. Compare Ex. D-502 with D-510.

374. Both products are in the same dosage form, i.e., a cream. Ex. D-102, D-502, D-510; Ex. G-12-8, Dr. Kanof, T-1250; Dr. Rhodes, T-1048. (A)

375. The Premo betamethasone valerate product is manufactured in accordance with good GMP's. Dr. Rhodes, T-1048-1049, Ex. D-102. (NDBI)

376. The inactive ingredients in the Premo betamethasone valerate product are generally recognized as safe and suitable. Dr. Rhodes, T-1204; Dr. Crout, T-274; Dr. Cabana, T-634-635. (NDBI)

377. Premo's betamethasone valerate product meets the applicable standards of the official compendium, the National Formulary. Dr. Rhodes, T-1209. (NDBI)

378. The quality control procedures instituted by Premo for the production of its Beta Val product are adequate and consistent with current pharmaceutical practice in the production of quality pharmaceutical products. Dr. Rhodes, T-1408-1409. (NDBI)

379. Notwithstanding the testimony of Dr. Rhodes, T-1221, I cannot say that Premo's betamethasone valerate does not present a medically significant bioavailability problem. As the Government notes, there are indications from the differences in the formulations of these products that they may differ in bioavailability.

a) Betamethasone valerate is a topically applied preparation intended for local therapeutic effect. Dr. Rhodes, T-1224. (A)

b) FDA's regulations, 21 C.F.R. § 320.22(b)(2), waive the requirement for submission of in vivo bioavailability data for such a product. Dr. Rhodes, T-1224. (A)

c) There is no documented evidence of therapeutic failure with betamethasone valerate. Dr. Cabana, T-613. (A)

380. I cannot find, as defendants request, that Premo's betamethasone valerate product is of equivalent safety and effectiveness to Schering's Valisone product. See Finding of Fact 379, supra.

381. The clinical trial conducted for Premo by Norman P. Kanof, M.D., Dr. Kanof, T-1232-1234; Ex. D-79, is flawed in design, has not been completed, and has not been subjected to any statistical evaluation.

382. Dr. Kanof has been conducting, for the past several months, a clinical trial to compare the safety and effectiveness of Premo's Beta Val to Schering's Valisone. T-1237; Ex. D-57. (A)

383. At the time of his testimony, Dr. Kanof had completed the study with respect to 24 subjects. T-1239; Ex. D-57. (A)

384. To the present time, Dr. Kanof has noted no difference between the two products in any of the patients that he has treated. The two products have been equally effective, and no side effects at all have been noted. Dr. Kanof, T-1240, 1244; Ex. D-57. (A)

385. Through March 12, 1980, Premo had distributed 342,000 tubes of Beta Val in interstate commerce. Stipulation, Ex. G-64. (A)

386. There have been no confirmed reports of therapeutic failure or adverse effect with the Premo product. Seymour Blackman, T-2218. (A)

387. Allopurinol is used clinically to lower uric acid levels in the blood by the inhibition of the enzyme zanthine oxidase. Its primary therapeutic use is in the treatment of gout. See Ex. D-501, D-509. (A)

388. Premo's allopurinol tablets, 100 and 300 mg, purport to contain the same amount of the same active ingredient as does Burroughs-Welcome's Zyloprim brand of allopurinol 100 and 300 mg tablets. Ex. D-103, D-104, D-501, D-509, D-43; Ex. G-12-3. (A)

389. The Premo and Burroughs-Welcome allopurinol products are in the same dosage form. Ex. D-103, D-104, D-501, D-509, D-43; Ex. G-12-3, Dr. Rhodes, T-1274, 1281. (A)

390. The Premo allopurinol products are manufactured in accordance with good GMP's. Dr. Rhodes, T-1269, Ex. D-103, D-104. (NDBI)

391. Premo's quality control procedures for the production of its allopurinol products meet with current good pharmaceutical practice in the production of finished pharmaceutical products. Dr. Rhodes, T-1271; Ex. D-103, D-104. (A)

392. The inactive ingredients utilized by Premo in its allopurinol product are generally recognized by qualified experts as safe and suitable for the production of finished pharmaceuticals. Dr. Rhodes, T-1205; See Dr. Cabana, T-634-635; Ex. D-103, D-104, D-43; Ex. G-12-3. (NDBI)

393. Mr. Kahan, of Fittelson Laboratories, New York, is a subcontractor of U.S. Testing and assayed the Premo and Zyloprim allopurinol products, 100 and 300 mg, for quantities of designated components. Kahan Dep. at 72, 75. (A)

394. Fittelson Laboratories is qualified to do this analysis. Kahan Dep. at 7-11, 92 (Ex. D-67B). (A)

Premo's Proposed Findings 87(c) to (h), submitted under seal, are now treated as follows:

395. 87(c) is found as submitted except that the first line should read, "Fettelson Laboratories aimed to determine the quantity —."

396. 87(d) is found as submitted except that following "77" in the last line, the words "except to the extent that it requires an assumption concerning the amount of magnesium in the magnesium stearate used."

397. 87(e) is found as submitted except, on the third line, the word "generally" should be inserted between "is" and "acceptable."

398. 87(f) is found as submitted except, on the second line, the word "generally" should be inserted between "is" and "acceptable."

399. 87(g) is found as submitted.

400. 87(h) is rejected.

401. All methods discussed in subparagraphs d-h above were followed by Fittelson Laboratories. (A)

402. The results obtained by Fittelson Laboratories in Ex. D-305 are the same as those submitted by U.S. Testing to Premo, Ex. D-43. Compare Ex. D-305 with D-43. (A)

Premo's Proposed Findings 89-90, submitted under seal, are now treated as follows:

403. 89 is found as submitted, except that there is added, on the last line, "However, as Dr. Shangraw testified, there are a number of potential differences between the active ingredients used in the two products that could be important with respect to the effects that these products would have in patients to whom they were administered."

404. 90 is found as submitted, except there is to be added the same material added to 89 in Finding of Fact 403, supra.

405. The Premo allopurinol products comply with the USP standards. Dr. Rhodes, T-1205, 1406-1407; Ex. D-103, D-104. (NDBI)

406. It is undisputed that allopurinol has a high therapeutic ratio. Dr. Bloom, T-391; Dr. Greene, T-1878.

407. Physicians do not monitor patients for blood levels of allopurinol. Dr. Bloom, T-391. (A)

408. It is not disputed that, if the Premo product were modestly more bioavailable than Burroughs-Welcome's, there would be no great medical consequences, and no hazard. It would not produce any toxic effects. Dr. Crout, T-47; Dr. Bloom T-392, 449; Dr. Greene, T-1878, 1880. (A)

409. There is no documented evidence of therapeutic failure of allopurinol. Dr. Cabana, T-606. (A)

410. The government witnesses testified that if the Premo product were significantly less bioavailable than Zyloprim, then the drug could be potentially harmful when used as an adjunct to cancer chemotherapy for the purpose of reducing serum uric acid levels. Dr. Rhodes admitted this is a "possibility." Dr. Rhodes T-1635. (A)

411. Premo has commissioned a bioavailability study. The results of a four-patient bioavailability pilot study have been reported and while the results thus far indicate that the two products are bioequivalent, the study is incomplete and I gave it no weight. Dr. Rhodes, T-1290-1291; Dr. Greene, T-1882-1888. Ex. D-85, D-152, D-152A, D-152B, D-152C.

412. Premo has sold 19.6 million allopurinol tablets as of March 19, 1980. Stipulation, Ex. G-64. (A)

413. There have been no confirmed reports of therapeutic failure with respect to allopurinol. Seymour N. Blackman, T-2217-2218; Dr. Crout, Affidavit, Ex. G-4. (A)

414. Hydroxyzine hydrochloride is indicated for use in the management of anxiety, tension, and psychomotor agitation in conditions of emotional stress. Ex. D-500, D-508. (A)

415. Premo's hydroxyzine hydrochloride purports to contain the same amount of the same active ingredient as does Pfizer's Atarax brand of hydroxyzine hydrochloride. Ex. D-105, D-106, D-107; Ex. G-12-6; Ex. D-501, D-509; Dr. Rhodes, T-1353-1354.

416. The Premo and Pfizer products are both sold in the same dosage form, i.e., 10, 20, and 50 mg tablets. Ex. D-105, D-106, D-107; Ex. G-12-6; Ex. D-501, D-509, D-76. (A)

417. Both products are labeled for the conditions of use of the drug. Compare Ex. D-501 with Ex. D-509.

418. The Premo hydroxyzine hydrochloride drug is manufactured in accordance with current good manufacturing practice. Dr. Rhodes, T-1353-1354. (NDBI)

419. Premo's hydroxyzine hydrochloride products meet the compendial standards of the USP. Dr. Rhodes, T-1210, 1353. (NDBI)

420. The inactive ingredients in the Premo hydroxyzine hydrochloride product are generally recognized as being safe and suitable for the production of finished pharmaceutical product. Dr. Rhodes, T-1204; Dr. Crout, T-277; Dr. Cabana, T-634-635. (NDBI)

421. Premo's quality control standards for the production of its hydroxyzine hydrochloride product are adequate and meet current requirements for the production of finished pharmaceutical products. Dr. Rhodes, T-1353-1354, Ex. D-105, D-107. (NDBI)

422. Premo's hydroxyzine hydrochloride product does not present any form of bioequivalence problem, much less a medically equivalent one. Dr. Cabana, T-588; Dr. Rhodes T-1136. (A)

a) It is unnecessary, according to the Food and Drug Administration, to have in vivo bioavailability data for hydroxyzine hydrochloride. (A)

b) FDA has advised Premo that based on the dissolution data for its 25 mg dosage, there is no need for in vivo bioavailability data to establish therapeutic equivalence. Ex. D-555. (A)

423. Since April of 1978, Premo has sold 34.6 million hydroxyzine hydrochloride tablets. Stipulation, Ex. G-604. (A)

424. There have been no confirmed reports of adverse reaction or therapeutic failure. Seymour N. Blackman, T-2217-2218. Mr. Blackman, however, stated that there was a report of therapeutic failure with this drug (Tr. at 2218) and there was no indication that he made any effort to confirm whether or not this was a true report.

425. Hydroxyzine Pamoate is indicated for the management of anxiety, tension, and psychomotor agitation in conditions of emotional stress. Ex. D-504, D-512. (A)

426. Premo's hydroxyzine pamoate capsules, 25, 50, and 100 mg, purport to contain the same amount of the same active ingredient as the hydroxyzine pamoate capsules, 25, 50, and 100 mg, marketed by Pfizer under the trade name Vistaril. Ex. D-108, D-109, D-109A, D-512; Ex. G-12-5.

427. Both products are in the same dosage form, i.e., a capsule. T-1357; Ex. D-108, D-109, D-109E, D-504, D-512; Ex. G-12-5. (A)

428. The Premo hydroxyzine pamoate product is manufactured in accordance with good GMP's. Dr. Rhodes, T-1357. (NDBI)

429. The inactive ingredients of the Premo hydroxyzine pamoate capsules are generally recognized as safe and suitable. Dr. Rhodes, T-1204; Dr. Crout, T-278; Dr. Cabana, T-634-635. (NDBI)

430. Premo's hydroxyzine pamoate capsules meet the requirements of the USP. Dr. Rhodes, T-1211. (NDBI)

431. The Premo and Pfizer hydroxyzine pamoate capsule products are similarly labeled for use. Compare Ex. D-504 with Ex. D-512.

432. The Premo hydroxyzine pamoate capsules are produced in accordance with quality control procedures which meet current standards for the manufacture of finished pharmaceutical products. Dr. Rhodes, T-1356; Ex. D-108, D-109, D-109A. (NDBI)

433. Hydroxyzine pamoate is labeled to imply that it is equivalent to hydroxyzine hydrochloride. Dr. Cabana, T-606. (A)

434. FDA is unaware of any data demonstrating the equivalence of hydroxyzine pamoate to hydroxyzine hydrochloride. (A)

435. Such studies are required from Pfizer the innovator, as a condition for its continued approval of the drug. This requirement was published in the Federal Register of February 2, 1979. Dr. Cabana, T-610. 44 Fed.Reg. 6780, attached to Affidavit of Dr. Crout, Ex. G-4. (A)

436. As of April 11, 1980, the study has not been completed by Pfizer. Dr. Cabana was not even sure that it had been started. Dr. Cabana, T-607.

437. FDA has not taken any action against Pfizer or its Vistaril brand of hydroxyzine pamoate capsules. Dr. Cabana, T-608. Thus, Pfizer's hydroxyzine pamoate remains on the market, even though the innovator has not completed a bioavailability study which FDA says is necessary. Dr. Cabana, T-607. Dr. Cabana explained the practical problems that require the agency to leave the innovator product on the market. The statute would require a hearing to remove it since the product has been approved. See 21 U.S.C. § 355(e). If Pfizer does not complete the necessary testing satisfactorily, action can be taken against its product.

438. Premo has sold over 12.29 million dosage units of its hydroxyzine pamoate product and there have been no reports of adverse reaction or complaints. Seymour Blackman, T-2217-2218. (A)

439. Triamterene with hydrochlorothiazide is indicated for treatment of edema and hypertension. Ex. D-506, D-514. (A)

440. The Premo Triamthiazide product purports to contain the same amounts of the same two active ingredients as does the innovator product, Smith, Kline and French's Dyazide brand of triamterene with hydrochlorothiazide capsules. Dr. Crout, T-25; Dr. Rhodes, T-1300, 1405; Ex. D-111-130; Ex. G-12-1; Ex. D-506, D-514, D-48.

441. The Premo product is in the same dosage form, i.e., a capsule, as the Smith, Kline and French product. Ex. D-111-130; Ex. G-12-1; Ex. D-506, D-514, D-48. (A)

442. The Premo Triamthiazide product is manufactured in accordance with good manufacturing practice. Dr. Rhodes, T-1293-1294. (NDBI)

443. The inactive ingredients contained in the Premo product are safe and suitable for use in the production of finished pharmaceutical products. Dr. Rhodes, T-1198. See Dr. Crout, T-256; Dr. Cabana, T-634-635. (NDBI)

Premo's Proposed Findings 137(a)-(g) submitted under seal, are now treated as follows:

444. 137(a), (b), (e), and (g) are found as submitted.

445. 137(c) is found as submitted except that there is to be added, "I do not find, however, that the results obtained are necessarily reliable."

446. 137(d) is rejected as stated.

447. 137(f) is rejected as stated.

Premo's Proposed Findings 138(a)-(c), submitted under seal, are now treated as follows:

448. 138 is found as submitted except that Dr. Crout, upon whose testimony this proposed finding is based, made it clear that he lacked expertise on drug composition and deferred to other witnesses; Dr. Shangraw noted that there is a difference in the formulas in terms of the kind of lactose used.

449. 138(a), first sentence, is found as submitted. With respect to the second sentence, Dr. Crout emphasized that he was not qualified to determine whether the difference would or should cause concern over bioavailability of the Premo product.

450. 138(b) is found as submitted except that there is to be added that Dr. Rhodes suggested that the magnesium stearate difference would lead to a slower dissolution of the Premo product.

451. 138(c) is rejected on the basis of the Shangraw reference cited in the Critique of the United States to Premo's Proposed Finding 138(c).

452. Both the Premo Triamthiazide and Smith, Kline and French Dyazide products are labeled for use in the treatment of hypertension and edema. Compare Ex. D-506 with Ex. D-514.

453. The active ingredients in the Premo Triamthiazide product comply with the requirements of the USP. Dr. Rhodes, T-1405. (NDBI)

454. The quality control procedures utilized by Premo in the production of its Triamthiazide product are adequate and meet current requirements for the production of finished pharmaceutical product. Dr. Rhodes, T-1405. (NDBI)

455. The record contains a dispute as to whether or not the drug triamterene with hydrochlorothiazide presents a medically significant bioavailability problem which requires some understanding of the pharmacology of the drugs in order to resolve. Compare, e.g., Dr. Bloom, T-425 with Dr. Cabana, T-484. (NDBI)

456. The drug hydrochlorothiazide is a known effective diuretic drug. Among its pharmacological properties is an increased excretion of potassium from the body. Ballantyne Dep. at 106 (Ex. D-49); Dr. Rhodes, T-1621; Dr. Crout, T-33, 34. (A)

457. The triamterene component of the product, although having some diuretic effect, is primarily present for its potassium-sparing effect in conjunction with use of the hydrochlorothiazide. Dr. Crout, T-33-34; Dr. Cabana, T-155; Dr. Rhodes, T-1141-1146; Dr. Cortell, T-155; and Dr. Bloom T-405 (Ex. D-78).

458. The Government contends that the Premo product presents a medically significant bioavailability problem, in that, if the Premo product has significantly greater bioavailability than the Smith, Kline and French product, the resulting increased absorption of triamterene, with its potassium-sparing effect, may lead to a build-up of potassium in patients, thus leading to the condition known as hyperkalemia, or increased potassium in the blood. Dr. Crout, T-38; Dr. Cabana, T-593; Dr. Cortell, T-155-156.

a) According to the government, the potential for hyperkalemia from an excess of triamterene does not exist in the majority of the population. Triamterene is contraindicated for patients with severe progressive kidney disease. Dr. Crout, T-96. (A)

b) As a general principle, for patients with adequate urinary output, hyperkalemia will rarely occur in patients on triamterene. Dr. Crout, T-100. (A)

459. At the time of trial, however, FDA has no information, from bioavailability testing, that Premo's marketed Triamthiazide product is more bioavailable than Dyazide. Dr. Crout, T-93. I cannot, however, reject the contention of the United States that there is, however, reason to believe based upon the differences in the formulations of the drug that Premo's product would be, in fact, more bioavailable.

460. Both drugs have a flat dose-response curve. With a flat dose-response, as the dose is increased, there is no greater therapeutic effect. Dr. Barr, T-982, 935; Dr. Rhodes, T-1039, 1139. However, this product has flat dose-response curves only for its therapeutic effects, not for the toxic or adverse effects that it may cause ( see Tr. at 935-36 (Dr. Barr)).

461. Where there is a flat dose-response curve, even if there is a difference in bioavailability, it is meaningless in a clinical sense. Dr. Barr, T-982, Dr. Rhodes, T-1039, 1140. However, this does not mean that a flat dose-response curve for therapeutic effect by itself would mean that a difference in bioavailability would be meaningless in a clinical sense where there might be a steeper dose response curve for other effects ( see Tr. at 935-36 (Dr. Barr)).

462. Dr. Cabana testified that the product presented a medically significant bioequivalence problem. Dr. Cabana, T-597; Ex. D-140. (A) Notwithstanding defendants' attempts to discredit his testimony ( see Defendants' Proposed Finding 146(d)), I credit his testimony.

463. Dyazide is not a drug where physicians precisely monitor blood levels of the drugs in their patients. Dr. Cortell, T-171; Ballantyne Dep. at 85-86 (Ex. D-71B); Dr. Bloom, T-391.

464. There is no well documented evidence of therapeutic failure with the drug hydrochlorothiazide. Dr. Cabana, T-593. (A)

465. None of the government's witnesses knew how much additional triamterene delivered to the patient would be likely to cause hyperkalemia in the patient with mild to moderate kidney disease. Dr. Crout, T-108; Dr. Cortell, T-172. (A) It is noted, however, that as Dr. Crout testified, one would not be expected to have this information since a test that involved inducing hyperkalemia in this type of patient would be unlikely (Tr. at 108); and Dr. Cortell testified that a small change in bioavailability of triamterene might affect the serum potassium in patients with kidney disease (Tr. at 171-72).

466. If the Premo product were to present a known medically significant bioavailability problem, the safety and effectiveness of the Premo product as compared to Dyazide could be established on the basis of in vivo bioavailability studies. See Dr. Cabana, T-478. (A)

467. Premo has commissioned an in vivo bioavailability study of its Triamthiazide product comparing it with the reference product, Dyazide. Dr. Rhodes, T-1405; Dr. Greene, T-1792-1799; Ex. D-83 Revised, Ex. D-84. (A)

468. Triamthiazide is a combination product. The in vivo bioavailability study, not yet complete, conducted on behalf of Premo analyzes blood levels of both the triamterene and hydrochlorothiazide components in sixteen subjects. Dr. Greene, T-1853. (A)

469. Two completely different assays were performed and two different sets of data generated with respect to the two components, triamterene and hydrochlorothiazide. Dr. Greene, T-1853; Cohn Dep. 148 (Ex. D-66B); D-83 Revised, D-84. (A)

470. Ex. D-83 Revised (the study of the triamterene component) was revised by Dr. Greene and his colleague Bruce Birmingham from Ex. D-83 in order to 1) equalize the number of subjects in both treatment groups, since this is preferable from the standpoint of statistical analysis; and 2) eliminate subject 2 since subject 2 had two points missing at the peak of the curve and inclusion of subject 2 was inadvisable since it is difficult to give an honest and accurate estimate of the two sequential points. Dr. Greene, T-1813. (A)

471. In Ex. D-83 Revised, Dr. Greene and Mr. Birmingham used a random number generator to equalize the treatment groups and thereby eliminated subjects 3, 11, and 6. T-1813. (A)

472. Ex. D-83 Revised contains the following data:

a) The total area under the curve, expressed in nanograms per mil per 24 hours, was determined for each of the sixteen subjects. T-1816(A) Dr. Greene's calculations of area under the curve were made according to a computer program written by Dr. Greene which calculates the area under the curve using the trapezoidal rule. T-1814; Ex. D-83 Revised. (A)

b) An analysis of variance consists of statistical tests that are performed to determine whether the factors being considered, i.e., the concentrations at various discrete time points, AUS, C-max and T-max, are statistically different. T-1817. Dr. Greene performed an analysis of variance for the triamterene bioavailability data. T-1817; Ex. D-83 Revised at 2-26. (A)

c) The elementary statistics, i.e., the average values, the variance, standard deviation, minimum and maximum values of the triamterene component for both the SKF and Premo products were calculated by Dr. Greene. T-1817; Ex. D-83 Revised at 28-38. (A)

d) Dr. Greene prepared a "summary of results" which summarizes the elementary statistics, and also performed a summary of the statistical tests performed on all the time points, peak height, time to peak and area under the curve. T-1817. (A)

473. Dr. Greene prepared his analysis in Ex. D-83 Revised from raw data on triamterene blood levels provided by National Medical Services. T-1821; Ex. D-356. (A)

a) National Medical Services assayed the triamterene component using a spectrofluorometric technique. Dr. Greene, T-1902; Cohn Dep. at 69 (Ex. D-66B); Ex. D-201. (A)

b) Dr. Greene is familiar with the method used by Dr. Cohn, has reviewed the assay, and finds it perfectly suitable. T-1798, 1903. (A)

c) According to the literature, the extraction techniques used by National Medical Services separates out the metabolyte and it will not be detected as triamterene. Dr. Greene, T-1903; Cohn Dep. at 112 (Ex. D-66B); Ex. D-219, D-220. (A)

474. Dr. Cohn of National Medical Services, obtained the blood specimens with triamterene from Clinical Concepts, Inc. D-353-355; Robillard Dep. 81-90 (Ex. D-68B); Cohn Dep. at 75-76 (Ex. D-66B). (A)

a) The protocol for the bioavailability test, D-349, specifies how the blood samples were to be collected. Robillard Dep. at 36, 46, 66-75 (Ex. D-68B). (A)

b) Clinical Concepts selected the subjects, administered the Premo and SKF triamterene hydrochlorothiazide products, and collected the blood in accordance with Ex. D-349 (the protocol) with the exception that the treatment groups were inadvertently unbalanced. Robillard Dep. at 68 (Ex. D-68B); Ex. D-351. (A)

c) There were adequate measures taken by Clinical Concepts to ensure that qualified subjects participated in the study. Robillard Dep. at 51-54 (Ex. D-68B). (A)

d) There were adequate measures taken by Clinical Concepts to ensure that the subjects followed the requirements for participation in the study. Robillard Dep. at 56-59 (Ex. D-68B); Ex. D-349. (A)

e) There were adequate measures taken by Clinical Concepts to ensure timely collection of blood at appropriate intervals. Robillard Dep. at 66-72 (Ex. D-68B); Ex. D-351. (A) However, wherever the error, at least one of the subjects, even in Dr. Greene's opinion, had to be dropped because of inadequate numbers of blood samples.

f) There were adequate measures taken by Clinical Concepts to ensure that the correct labels were affixed to all specimens. Robillard Dep. at 71-75 (Ex. D-68B); Ex. D-351. (A)

g) There were adequate measures taken by Clinical Concepts to ensure that the specimens were adequately stored and transported to National Medical Services. Robillard Dep. at 78-91 (Ex. D-68B); Ex. D-353-355. (A) See, however, Finding 474(e), supra.

475. Ex. D-84 (the study of the hydrochlorothiazide component) was revised by Dr. Greene. Dr. Greene, T-1856.

476. Findings 470-471 regarding the use of a random number generator apply with respect to Ex. D-84.

477. Ex. D-84 contains the same type of data analyses described in Finding 472(a) through (d) above. Dr. Greene, T-1851-1852.

478. Dr. Greene performed his analysis from the hydrochloride blood concentrations analyzed by National Medical Services. T-1856. (A)

a) A high pressure liquid chromatographic (HPLC) assay performed by Dr. Cohn, of National Medical Services, was utilized to determine the quantity of hydrochlorothiazide in the samples. Dr. Greene, T-1853-1854; Cohn Dep. at 116-121 (Ex. D-66B). (A)

b) Dr. Greene is familiar with this assay technique; Dr. Greene examined the raw data (the chromatograms, the laboratory notebook, and the depositions of Dr. Cohn). T-1853-54. (A)

c) The system response is linear. Cohn Dep. at 205 (Ex. D-66B)

479. Dr. Cohn of National Medical Services obtained blood specimens with hydrochlorothiazide from Clinical Concepts, Inc. D-353-355; Robillard Dep. 82-90 (Ex. D-68B). (A)

a) The Premo product reaches T-max a half an hour before the SKF product does with respect to the triamterene component. Dr. Greene, T-1834; Ex. D-148. (A)

b) Furthermore, statistical analysis performed on the data to determine if the actual differences reflected in the means of C-max, T-max, and AUC establish that the differences are not statistically significant. T-1837; Ex. D-83 Revised at 4; Ex. D-148, D-148A. However, a lack of statistical significance can be, as it may well be with this test, simply a result of the insensitivity of the test to detect a difference if there is one.

c) With respect to the hydrochlorothiazide component, an analysis of variance for the C-max parameter of the hydrochlorothiazide component shows a statistically significant difference. (A)

480. In response to FDA raising the possibility that the Premo product might raise serum potassium levels and therefore cause hyperkalemia, Premo commissioned a comparative potassium sparing study. Dr. Rhodes, T-1324.

481. The duration of treatment in the potassium sparing study was a total of three days for each product. Ex. D-49. (A)

482. There is a scientific dispute as to whether the Premo bioavailability studies had sufficient "power" to detect a difference between the two products. Compare Dr. Greene, T-1838 with Dr. Barr, T-945, 951. (A)

483. As indicated, see Finding of Fact 257, supra, a power analysis is a statistical test which is used to determine the sensitivity of a test to detect certain differences between any given number of groups. The power of the test is expressed as a percentage which provides a level of confidence at which a particular test can detect a given difference between two or more things being measured. Dr. Greene, T-1846; see Dr. Cabana, T-502; Dr. Barr, T-912. (A)

484. According to a power analysis conducted by FDA, the power of the Premo test, Ex. D-83 Revised, to detect a 20% difference is 20 or 30%. Dr. Barr, T-940, 147; Ex. G-208. (A)

485. Premo sold 24 million capsules of its Triamthiazide product between November 14, 1979, and December 17, 1979. Stipulation G-64. (A)

485a. Chlorthalidone tablets, 50 mg and 100 mg, are indicated for the management of hypertension and as an adjunctive therapy in edema. D-505, 513. (A)

486. Premo's chlorthalidone tablets, 50 mg and 100 mg, purport to contain the same amount of the same active ingredients as does the USV Hygroton brand of chlorthalidone 50 mg and 100 mg tablets, and both products are labeled for their intended conditions of use. Dr. Rhodes, T-1339; Ex. D-100-101; Ex. G-12-2; Ex. D-505, D-513.

487. The two products are in the same dosage form, i.e., a tablet. Ex. D-100-101; Ex. G-12-2; Ex. D-505, D-513. (A)

488. The Premo chlorthalidone products are manufactured in accordance with good manufacturing practice. Dr. Rhodes, T-1337; Ex. D-100-101. (NDBI)

489. The inactive ingredients in the Premo chlorthalidone product, 50 and 100 mg, are generally recognized as safe and suitable for use in the production of finished pharmaceutical product. Dr. Rhodes, T-1201; Dr. Cabana, T-634-635; and Dr. Crout, T-272. (NDBI)

490. The Premo chlorthalidone tablets meet the requirements of the United States Pharmacopeia (Dr. Rhodes, T-1209), as well as additional inhouse controls over and above the requirements of the USP. Dr. Rhodes, T-1338. (NDBI)

491. The quality control procedures utilized by Premo in the production of its chlorthalidone tablets are perfectly acceptable and meet current good standards for the production of any pharmaceuticals. Dr. Rhodes, T-1337.

492. The record reveals a dispute as to whether or not chlorthalidone presents a medically significant bioavailability problem. Dr. Rhodes testified that such a problem with chlorthalidone is "highly improbable." Dr. Rhodes, T-1215. But see Dr. Cabana, T-602. Government experts strongly asserted that there would be such a problem.

493. Physicians do not monitor patients for blood levels of chlorthalidone. Dr. Bloom, T-391; Dr. Barr, T-1004, Ex. D-154B (USP Dispensing Information on chlorthalidone). (A)

494. Physicians do not monitor patients for blood levels of chlorthalidone. Dr. Bloom, T-391; Dr. Barr, T-1004, Ex. D-154B (USP Dispensing Information on chlorthalidone). (A)

495. Chlorthalidone has a flat dose-response curve for therapeutic purposes. Dr. Barr, T-935. (A) However, it should be noted that there is a rising dose response curve for chlorthalidone for its toxic reactions (Tr. at 935 (Dr. Barr)).

496. On the other hand, Dr. Cabana testified that based on data made available to the Food and Drug Administration concerning some other product other than Premo, there was data to suggest a difference of 60-70% in bioavailability. T-602. (A)

497. Also, based on the Tweedale and Matterson studies, Ex. D-144, D-145, a difference of less than two-fold in chlorthalidone appears to affect potassium levels. Dr. Cabana, T-618. (A) However, Dr. Cabana also made it clear that the literature indicates that a difference of from 50 to 75 or 100 milligrams can lead not only to an effect on potassium levels but also to hypokalemia (Tr. at 618).

498. Both the Premo and USV products are labeled for use. Compare D-505 with D-513.

499. Premo has undertaken an in vivo bioavailability study for its chlorthalidone products, comparing the blood levels produced by the Premo product to the blood levels produced in the same subjects by the USV product. Ex. D-39, D-39A. (A)

a) The chlorthalidone study was conducted as a three-way crossover. Robillard Dep. at 123 (Ex. D-68B). A three-way crossover is when all the subjects are divided into three groups:

(1) The first group would receive drugs "A," "B," and "C," in the order A, B, and C; the second group would receive the drugs in the order B, C, and A; and the third group would receive C, A, and B. Dr. Greene, T-1743.

(2) The drugs studied in the Premo three-way crossover chlorthalidone study were the Premo 50 mg, Premo 100 mg, and USV Hygroton 100 mg. Dr. Greene, T 1743; Robillard Dep. at 123 (Ex. D-68B).

(3) The advantage of a three-way crossover is that external factors would affect all three of the products equally instead of biasing it toward one product. Dr. Greene, T-1744.

b) Blood samples were collected at discrete intervals after the administration of the drug. The samples were then collected, stored, and assayed. Dr. Greene, T-1745. (A)

c) The raw data concerning the measurement from these three products in the blood is contained in D-39. Dr. Greene, T-1743; Ex. D-39. (A)

d) The same raw data contained in D-39 is also contained in Ex. D-39A. Dr. Greene, T-1747. (A)

e) As originally reported, the analysis of the data in Ex. D-39 contained several errors as pointed out by Dr. Cabana, including errors in transposing the raw data into the computer for statistical analysis. Dr. Cabana, T-507-510. (A)

f) The data was also subject to criticism, in Dr. Cabana's opinion, because a number of the subjects had missing values ( i.e., the samples were lost or broken and therefore not analyzed for certain time periods) where the value had been substituted. Dr. Cabana, T-505. (A)

g) Ex. D-39A, prepared by Dr. Greene, re-analyzes the data to correct for the errors previously pointed out by Dr. Cabana. Dr. Greene, T-1746, 2039. (A) He eliminated only, however, those subjects that were missing more than two data points (Tr. at 2039). He, of course, would be unable to correct the study and its lack of data points for subjects if they had one or two missing data points. He further did not report any corrected results for the urine data, which contained many errors.

(1) Dr. Greene took the original raw data in Ex. D 39 and performed his analysis. Dr. Greene, T-1747. (A)

(2) Dr. Greene corrected the transposition errors. Dr. Greene, T-1749. (A)

(3) Dr. Greene also eliminated a subject if more than two data points were missing. T-1751. (A)

(4) If two or less points were missing, he estimated the value using a standard pharmacokinetic techniques. T-1752, 2041. (A)

h) Dr. Greene then placed the subjects into three balanced groups of five subjects each. Dr. Greene, T-1754. (A)

i) Dr. Greene, in his analysis attempted to compensate for the missing data points and corrected arithmetic errors in Ex. D-39. Dr. Barr, T-966, 971. (A)

j) Using a computer program, Dr. Greene conducted what is known as an analysis of variance, a standard technique used to determine whether two or more groups are statistically the same or are statistically different. (A)

k) The FDA's own test protocol for a bioavailability study for the drug Digoxin (a drug with very important, medically significant bioavailability problems which has a very low therapeutic ratio, Dr. Greene, T-1778; see Finding 364 above) shows that the FDA test has only a 50% power to detect a 20% difference. Dr. Greene, T-1778, D-400. (A)

500. Over 26 million Premo chlorthalidone tablets have been marketed since June 1979. Stipulation Ex. G-64. (A)

501. There have been no confirmed reports of adverse reaction or therapeutic failure received by Premo (Seymour Blackman, T-2217-2218), or FDA. Ex. G-4, Affidavit of Dr. Crout. (A)

502. The Premo Doxine product purports to contain the same amount of the same active ingredient as does the Richardson-Merrell-Bendectin product. Dr. Rhodes, T-1410; Ex. D-110, D-503, D-511; Ex. G-12-7.

503. The drugs are available in the same dosage form. Ex. D-110, D-503, D-511; Ex. G-12-7. (A)

504. The Premo product is manufactured in accordance with good manufacturing practice. Dr. Rhodes, T-1360. (NDBI)

505. The inactive ingredients utilized by Premo in the production of its doxine product are generally recognized as safe and suitable. Dr. Rhodes, T-1205, 1410; Dr. Cabana, T-634-635; Dr. Crout, T-280. (NDBI)

506. The active ingredients used in the Premo product meet compendial standards. Dr. Rhodes, T-1211. (NDBI)

507. The quality control procedures utilized by Premo in the production of its product are appropriate and acceptable in terms of current pharmaceutical practice in the production of finished pharmaceuticals. Dr. Rhodes, T-1410. (NDBI)

508. The Premo and Bendectin products are both labeled for treatment of nausea in pregnancy. Compare Ex. D-503 with Ex. D-511.

509. Both pyridoxine hydrochloride and doxylamine succinate are available over-the-counter in dosages much higher than those contained in Doxine and Bendectin. Dr. Cabana, T-583; Dr. Rhodes, T-1226.

510. 14.58 million tablets of the Premo Doxine product have been distributed in interstate commerce since September, 1978. Stipulation, Ex. G-64 at ¶ 30. (A)

511. There have been no confirmed reports of adverse reaction or therapeutic failure received by Premo (Seymour Blackman, T-2217-2218), or by the FDA (Ex. G-4, Affidavit of Dr. Crout). (A)

512. Trifluoperazine is indicated as effective for use in the management of the manifestations of psychotic disorders. Ex. D-507, D-515. (A)

513. Both the Premo trifluoperazine 2 and 5 mg tablets purport to contain the same amount of the same active ingredient as does the Smith, Kline and French Stelazine brand of 2 and 5 mg trifluoperazine. Ex. D-98, D-99; Ex. G-12-4.

514. The two products are available in the same dosage form. Ex. D-98, D-99; Ex. G-12-4; Ex. D-507, D-515. (A)

515. The Premo product is manufactured in accordance with current good manufacturing practice. Dr. Rhodes, T-1380; Ex. D-98, D-99. Dr. Rhodes' opinion that the Premo trifluoperazine product is produced in accordance with current GMP's is based upon his inspection of Ex. D-98, D-99 and his visit to the Premo plant. T-1380. (NDBI)

516. The inactive ingredients utilized by Premo in the Production of its trifluoperazine products are generally recognized as safe and suitable. Dr. Rhodes, T-1205; Dr. Crout, T-282; Dr. Cabana, T-634-635. (NDBI)

517. The trifluoperazine product produced by Premo meets the official compendial requirements. Dr. Rhodes, T-1212, 1380; Ex. 98, 99. (NDBI)

518. The quality control procedures utilized by Premo in the production of its trifluoperazine products are adequate and acceptable and meet current standards for the production of finished pharmaceutical products. Dr. Rhodes, T-1380-1383; Ex. D-98, D-99. (NDBI)

519. a) According to Dr. Seife, Director of the Division of Generic Drug Monographs at FDA, there is no adequate methodology for testing the bioavailability of trifluoperazine. Seife Dep. at 52 (Ex. 81B). Dr. Seife is, however, not an expert in bioavailability, as he stated (P. Ex. 69-A, page 1; P. Ex. 69-B at 75-76 (Deposition of Dr. Seife)). Thus his conclusions concerning the existence of adequate methodology for the bioavailability testing of trifluoperazine are given little weight.

b) Dr. Perel is unaware of any known bioavailability or bioequivalence problem with trifluoperazine. T-375. (A) However, I find a clear potential for bioavailability problems with trifluoperazine.

520. Although Dr. Perel testified regarding his own methodology to test the bioavailability of trifluoperazine, he has never tested drugs by this method (T-374) and did not know whether FDA was aware of it. T-375. (A)

521. Dr. Cabana further testified that FDA has not approved any study on trifluoperazine with any of the methods that Dr. Perel testified about and has not approved any products utilizing the radio receptor method that Dr. Perel described. T-676-677.

522. For certain drugs with known medically significant bioequivalence problems, good medical practice requires the monitoring of the levels of the drug in the patient's blood as a precaution to make sure that the patient is not getting too much or too little of the drug. See Dr. Barr, T-999. For example, this is the case with digoxin, a drug with a low therapeutic ratio (Dr. Cabana, T-548), and with aspirin, which is a high risk bioavailability problem drug (Dr. Barr, T-999) where the physician is to monitor the aspirin levels in the patient's blood when the drug is used at high doses in the treatment of rheumatoid arthritis. Dr. Barr, T-999; Ex. D-153 (USPDI for aspirin). (A)

523. The USPDI 20 does not recommend that physicians monitor trifluoperazine blood levels of patients. Ex. D-154. (A)

524. The Agency has taken the position in this case that Premo must be enjoined from distributing its drugs because it is potentially dangerous to allow the marketing of drugs with "unknown" bioavailability. Ex. G-4 at ¶ 21 (Affidavit of Dr. Crout); Ex. G-2 (Affidavit of Dr. Cabana). The Agency and the Government's position is, of course, that Premo must be enjoined from distributing drugs without approval when the law requires those drugs to be approved. (A)

525. However, there are many hundreds of drug products available on the market, and lawfully permitted to be marketed by FDA, which are of unknown bioavailability. For example, several hundred products are still under review in the drug efficacy study implementation (DESI) program. Dr. Crout, T-19. (A)

526. FDA has no bioavailability data on these drugs. Byers Dep. at 30-32 (Ex. D-80B). (A)

527. Furthermore, Paragraph XIV drugs, DESI drugs, and drugs not yet initially evaluated, constitute a significant portion of prescription pharmaceuticals sold in the United States. Dr. Crout, T-186. (A)

528. FDA nonetheless permits marketing of these drugs including the innovator and generic versions without regulatory action. Dr. Crout, T-32. See Byers Dep. at 20, 22 (¶ XIV drugs); Byers Dep. at 33, 34 (less than effective drugs) (Ex. D-80B). The withdrawal of less than effective drugs can take six years or more from the time FDA proposes to withdraw the NDA. Byers Dep. at 37, 38 (Ex. D-80B); Dr. Crout, T-195. (A)

529. This is true even where the drug in question has known bioequivalence problems, such as in the case of phenylbutazone alkai. Dr. Crout, T-231, T-234. (A)

530. Aspirin is an OTC drug that is in the high risk category in terms of bioavailability problems, and good medical practice requires a monitoring of the patient's aspirin blood levels when used at high doses. FDA does not have bioavailability requirements for aspirin. Dr. Barr, T-999-1000. (A)

531. Over-the-counter drugs are being declared as old drugs by virtue of a monograph system. See 21 C.F.R. § 330. (A)

532. It is not disputed that drug products which comply with the monographs are considered by the agency not to be "new drugs." Drug products that conform to the monograph, not active ingredients, are declared to be generally recognized and that expert general recognition is of both safety and effectiveness.

533. No bioavailability data is needed to classify OTC drugs as generally recognized as safe and effective and not misbranded. Id., See also 21 C.F.R. § 330.10 (requirements for OTC antacid products do not include a demonstration of bioavailability). (A)

534. Drugs initially marketed prior to 1938 are "grandfathered" and not subject to the new-drug definition, hence, they are not new drugs, and may be marketed without NDA's. Byers Dep. at 13-16, 40 (Ex. D-80B). (A)

535. However, the FDA takes the position that any generic version of a pre-1938 drug may be marketed without preclearance, regardless of the time at which the generic version commenced marketing. Byers Dep. at 13-16, 40 (Ex. D-80B). (A)

536. The Agency does not have data from the manufacturers establishing the bioavailability of the generic versions. Byers Dep. at 13-16, 40 (Ex. D-80B). (A)

537. In certain circumstances, the Agency has established a bioequivalence requirement to ensure the therapeutic equivalence of the drugs regardless of whether the drug was a pre-1938 and, hence, not a new drug. See Byers Dep. at 18 (Ex. D-80B). An example is Digoxin. (A)

538. The establishment of a bioavailability requirement with respect to aspirin would be a reasonable approach for the Agency to take to ensure the therapeutic equivalency of all aspirin products, a drug presenting a high risk bioavailability problem when used at high doses. Dr. Barr, T-1001. (A)

539. No NDA's have been approved for solid dosage form post-1962 generic drug products. Dr. Crout, T-200. (A)

540. No bioequivalence requirements have been established for any of the eight drugs at issue, in the sense that FDA has not specified in a regulation the bioavailability tests to be used in evaluating these drugs.

541. There has been no judicial or binding administrative proceeding conducted pursuant to the terms of the Administrative Procedure Act, 5 U.S.C. § 552 et seq., resulting in a determination that any of the eight products at issue are new drugs. (A)

XVIII. Premo Has Failed To Show It Will Suffer Irreparable Harm if the Government's Injunction Is Granted.

542. The defendants' Proposed Findings 256-275 have been considered. In the light of all of the evidence, these proposed findings must be rejected, first, because Premo knew that it was treading a dangerous line in engaging in the activities discussed; and, second, the harm suffered by Premo if an injunction is granted does not outweigh the harm to plaintiff and the public if an injunction is denied.