AMERICAN HEALTH PRODUCTS CO., INC., General Nutrition Center, Inc., General Nutrition Corp., Melva Natural Products, Inc., Nature's Bounty, Inc., Nutrition Headquarters, Inc., Phoenix Laboratories, Inc., R-Kane Products, Inc., and Sunrise Chemical, Inc., Plaintiffs,
v.
Arthur Hull HAYES, Commissioner of Food and Drugs, and Food and Drug Administration, Defendants.
No. 82 Civ. 4492 (ADS).
United States District Court, S.D. New York.
November 14, 1983.*1499 Bass, Ullman & Lustigman, New York City, for plaintiffs; Lawrence H. Roth, Robert Ullman, Jacob Laufer, Steven Trost, New York City, of counsel.
Rudolph W. Giuliani, U.S. Atty., Southern District of New York, New York City, for defendants; Janis Farrell, Asst. U.S. Atty., New York City, Stephen D. Terman, Rockville, Md., Don Burley, of counsel.
MEMORANDUM OPINION AND ORDER
SOFAER, District Judge:
On July 1, 1982, the Food and Drug Administration ("FDA") announced its decision to classify as drugs under section 201(g)(1)(C) of the Federal Food, Drug, and Cosmetic Act (hereinafter the "Act"), a group of products generally known as "starchblockers." 21 U.S.C. § 321(g)(1)(C). A short time later plaintiffs, who manufacture starchblockers, brought this action for a declaratory judgment that their products are a "food" under the Act and therefore exempt from its premarketing approval requirements. The United States moved for a preliminary injunction against the sale of these products. After a hearing on the merits the government withdrew its motion and all the parties agreed to submit the substantive legal issue for final judgment. Subsequently, numerous actions from other districts have been consolidated with this one, on the premise that the decision here will control the results in those actions.[1]
*1500 The starchblockers under consideration here are derived from White Northern beans. The FDA maintains that the manufacturers make starchblocker tablets and powder by isolating and extracting the inhibitory protein from its natural source, as the labels of several of their products state. The manufacturers contend that a large proportion of the starch is simply removed from the beans in order to produce a flour with a high concentration of protein; starchblocker pills are then made by adding various binders and excipients. Whether characterized as extraction or concentration, it is undisputed that the process entails crushing the beans, heating them, removing their hulls, and then separating the starch constituents from the protein constituents by air-classification.
The beans contain a protein that inhibits the normal functioning of alpha-amylase, an enzyme produced by the pancreas. Alpha-amylase aids in the digestion of starch by breaking it down into glucose, which the body then absorbs and utilizes for energy. Plaintiffs market their product as an aid to weight reduction, claiming that, since the protein prevents the alpha-amylase from acting, starchblockers allow some starch to pass through the body undigested, enabling dieters to avoid calories. The protein thus functions as an antinutrient by interfering with the normal digestion, absorption, and utilization of starch.
Section 201(g)(1)(C) of the Act defines the term "drug" in part to mean "articles (other than food) intended to affect the structure or any function of the body of man or other animals." 21 U.S.C. § 321(g)(1)(C). The immediately preceding subsection defines "food" as "(1) articles used for food or drink for man or other animals, (2) chewing gum, and (3) articles used for components of any such article." Id. § 321(f). The starchblocker manufacturers concede that their products are intended to affect a human bodily function, but contend that the products are "food" and thus fall within the parenthetical exclusion of subsection (g)(1)(C). The government argues that the statutory definition of food does not encompass starchblockers, and seeks to enjoin their sale until the FDA approves them as a "drug." Unless a manufacturer can demonstrate that its product is "generally recognized as safe and effective," see id. § 321(p), classification as a drug requires the manufacturer to cease marketing its product until the FDA approves its new drug application. See id. § 355(a)[2].
*1501 In deciding whether the FDA's determination that starchblockers are drugs is "in accordance with law," 5 U.S.C. § 706(2)(A), the FDA's interpretation merits substantial deference, National Nutritional Foods Association v. Weinberger, 512 F.2d 688, 701-02 (2d Cir.) cert. denied, 423 U.S. 827, 96 S.Ct. 44, 46 L.Ed.2d 44 (1975). A court must only ensure that the agency's action was "governed by an intelligible statutory principle." National Nutritional Foods Association v. Mathews, 557 F.2d 325, 337 (2d Cir.1977). Here the government contends that subsection (g)(1)(C) contemplates dual classification; in addition to classifying as drugs all those products that affect bodily structure or functions and are not common foods, the Act is said also to classify as drugs within the part (C) definition even a "food" product, if it is sold with specific representations as to its physiological effects. The manufacturers argue, on the other hand, that the definition of "food" in subsection (f), which refers to common usage, governs the reach of the parenthetical exclusion in part (C). Therefore, they urge, if an article is a "food" under subsection (f), it cannot be regulated as a drug under subsection (g)(1)(C) regardless of any representations as to its structural or functional effects made in connection with its sale.
The government's contention is untenable. Though most sections of the Act countenance dual classification, no other contains a parenthetical like that Congress inserted in part (C). Ignoring that parenthetical would render meaningless the distinctions Congress has attempted to delineate. Nevertheless, the government is correct in claiming that starchblocker pills are a "drug" under the Act, because the pills are not a "food" in any sense cognizable under the statute. Accord Nutrilab, Inc. v. Schweiker, 713 F.2d 335 (7th Cir.1983).
I. Propriety of Dual Classification.
A. The Case Law
The government urges a construction of section 321(g)(1)(C) analogous to that of section 321(g)(1)(B). Under part (B) a product promoted for and therefore intended to be used for a therapeutic purposefor example, an orange sold as a cure for the common cold may be regulated as a drug although it is undoubtedly also a food. See, e.g., National Nutritional Foods Association v. Mathews, 557 F.2d at 334; United States v. Hohensee, 243 *1502 F.2d 367 (3d Cir.) cert. denied, 353 U.S. 976, 77 S.Ct. 1058, 1 L.Ed.2d 1136 (1957). This result is consistent with the language of the relevant definitions, because an orange sold in such a manner is both "used for food" and "intended for use in the diagnosis, cure, mitigation, treatment, or prevention of disease in man." But part (C) differs from its putative analogue in one glaring respect: it includes a parenthetical explicitly and unqualifiedly excluding "food" from its coverage.
The cases upon which the government relies for its contentions provide little if any support for the proposition that a food may be regulated under part (C). With two exceptions, they either rest on part (B), see, e.g., Hohensee, 243 F.2d at 368-70 & n. 1, or involve cosmetics, see, e.g., United States v. An Article ... Sudden Change, 409 F.2d 734, 739 (2d Cir.1969), and thus involve no explicit exception. While the court in United States v. Article of Drug BZD Silogen and Zymaferm, No. 74-L-136 (D.Neb. Sept. 1, 1976), declared intended use "[t]he touchstone under the statute" even though the manufacturer tried to bring its product within the part (C) parenthetical exclusion for food, it relied primarily on this same line of cosmetics cases and cases under part (B) without discussing the possibility of a distinction. And the court in United States v. "Vitasafe Formula M", 226 F.Supp. 266, 278 (D.N.J.1964), rev'd on other grounds, 345 F.2d 864 (3d Cir.1965), found that the "lipotropic factors" contained in a food article were "drugs within the meaning of 21 U.S.C. § 321(g) in that their only known use is to affect the structure and function of part of the body." While the court referred to part (C), regulation of the entire food article was not really at issue, since the statute would reach a determinate active ingredient within it. Cf. Premo Pharmaceutical Laboratories, Inc. v. United States, 629 F.2d 795, 799 (2d Cir.1980).
The government argues that even an explicit statutory exclusion such as the one in part (C) should not be read to preclude regulation where necessary to effectuate the broad protective purposes of the Act. Thus, according to the government, even were starchblockers classifiable as a food as a general matter, the manufacturers' labeling claims here reveal an intended use which brings them within the reach of the statute. The government finds support for its position in AMP Inc. v. Gardner, 389 F.2d 825 (2d Cir.), cert. denied, 393 U.S. 825, 89 S.Ct. 86, 21 L.Ed.2d 95 (1968), and United States v. An Article of Drug ... Bacto-Unidisk, 394 U.S. 784, 89 S.Ct. 1410, 22 L.Ed.2d 726 (1969).
Both AMP and Bacto-Unidisk addressed the superficially analogous question whether a product was properly classified a "drug" or excluded as a "device" under section 321 of the Act. Just as the parenthetical of subsection (g)(1)(C) excludes "food" from the coverage of that provision, the final clause of subsection (g)(1) excludes "devices" from the reach of the entire definition. The product at issue in AMP used nylon disk and threat to ligate severed blood vessels during surgery. The Second Circuit at least implicitly recognized that the product came within the literal terms of the definition of "device" in section 321(h), but it looked to the purpose of the statute to conclude that the product nevertheless was regulable as a drug within the terms of subsection (g). 389 F.2d at 826-30. In Bacto-Unidisk, the Supreme Court came to a similar conclusion with respect to a sensitivity disc used in laboratory tests to determine the proper antibiotic drug to administer to a given patient. Also guided by statutory purpose, and employing the analysis used in AMP, the Court explicitly rejected the contention that, since the sensitivity disc was "plainly a `device'" it was "by definition necessarily not a `drug.'" 394 U.S. at 798-99, 89 S.Ct. at 1418; see id. at 793-98, 89 S.Ct. at 1415-18. A careful reading of those opinions makes clear, however, that the courts were responding to specific indications in the legislative history that Congress intended the premarketing approval requirements for drugs to apply as well to devices.
The AMP court noted, first, that early versions of the Act, which underwent significant revision between its original introduction in 1933 and its final passage in *1503 1938, contained no separate definition of "device" but included the term within the definition of "drug", 389 F.2d at 827-28; second, that the parallel but independent definition of "device" was added in response to objections based only on the linguistic incongruity of denominating as drugs certain of the contraptions which the drafters intended the Act to cover, id. at 828-29; and third, that at the time the separate definition was added the distinction between drugs and devices had no practical significance because all the operative provisions of the bill applied equally to both, id. at 829. Some time after the drafters had separated the drug and device definitions, they added the premarketing approval provisions for drugs as a result of the deaths in late 1937 of nearly one hundred people from consumption of an untested drug. See 21 U.S.C. §§ 321(p), 355. See generally Report of the Secretary of Agriculture on Deaths Due to Elixir Sulfanilamide-Massengill, reprinted in 5 A Legislative History of the Food, Drug, and Cosmetic Act and Its Amendments ("Legislative History"). The AMP court observed, however, that these provisions employed the term "`new drug' ... without any attention to the fact that the distinction between `drug' and `device' had thus for the first time become important." 389 F.2d at 829. In light of this legislative history, the court held that the premarketing approval requirements applied to products which, though not strictly termed drugs, raised the dangers the approval process was intended to eliminate:
The exclusionary classification "devices" should, we think, be limited to such things as Congress expressly intended it to cover. The language of section 201(g) plainly permits calling AMP's nylon thread and disk, in their intended use, "drugs," and we hold that that is their appropriate classification.
Id. at 830. To have concluded otherwise would have converted a distinction grounded in concern for semantic integrity into a substantial, unintended limit on the reach of the "new drug" provisions, contrary to the principle that overly literal statutory interpretation should not defeat a congressional purpose clearly revealed in the legislative history. See, e.g., United States v. Public Utilities Comm'n of California, 345 U.S. 295, 315, 73 S.Ct. 706, 717, 97 L.Ed. 1020 (1953).
The Supreme Court in Bacto-Unidisk closely tracked the AMP court's analysis. In arriving at an identical conclusion, 394 U.S. at 798-801, 89 S.Ct. at 1418-19, it noted the common heritage of the definitions, the semantic objections which prompted their separation, the chronology of the various amendments, and the purpose of premarketing approval. Id. at 793-98, 89 S.Ct. at 1415-18.
The basis of the AMP and Bacto-Unidisk decisions in the particular development of the "drug"/"device" distinction deprives them of controlling significance here. The Act's definitions of "food" and "drug" have no such common heritage, and the Act throughout reflects an intention to treat these categories differently. The decisions do teach that this Act, like others designed to protect the public health, should be liberally construed to advance that purpose. See Whirlpool Corp. v. Marshall, 445 U.S. 1, 13, 100 S.Ct. 883, 891, 63 L.Ed.2d 154 (1980); United States v. Dotterweich, 320 U.S. 277, 64 S.Ct. 134, 88 L.Ed. 48 (1943). But they do not support the degree of definitional flexibility urged by the FDA here, for they neither change the rule that a statute's interpreter looks first to its language, see Santa Fe Industries, Inc. v. Green, 430 U.S. 462, 472, 97 S.Ct. 1292, 1300, 51 L.Ed.2d 480 (1977), nor eliminate the need to scrutinize the structure and history of the applicable provisions as carefully as did the AMP and Bacto-Unidisk courts themselves. See National Nutritional Foods Association v. Mathews, 557 F.2d 325, 336 (2d Cir.1977) (notwithstanding holding in Bacto-Unidisk, "when an FDA determination that an article is a `drug' is so directly in conflict with the statutory definition, it must be invalidated as arbitrary and capricious and not in accordance with law"); National Nutritional Foods Association v. FDA, 504 F.2d 761, 789 n. 35 (2d Cir.1974), cert. denied, *1504 420 U.S. 946, 95 S.Ct. 1326, 43 L.Ed.2d 424 (1975).
B. Language, Structure, and Legislative History of Section 321(g)(1)(C).
The most natural way to read the term "food" in the exclusion under part (C), as the manufacturers urge, is as a reference to the definition of that term in the immediately preceding subsection, which was added by Congress at the same time. However, the term would then refer most plausibly to that entire definition, not solely its first element, which concerns common usage. That the drafters chose to repeat the defined term in the first part of its own definition "articles used for food or drink for man or other animals" suggests only that they intended to include the everyday meaning of food as one component of the statutory meaning. Had they intended the parenthetical exclusion of part (C) to encompass only this everyday meaning, they could easily have repeated the words "used for food" appearing in subsection (f)(1), instead of employing the defined term, "food."
The second element of the subsection (f) definition of food is chewing gum. Consistent with its interpretive stance generally, the government argues by example that a chewing gum marketed as a laxative would be regulable as a drug under section 321(g)(1)(C). The manufacturers would presumably argue that the parenthetical excludes such a product unless some ingredient independently was classifiable as a drug.
The third element of the subsection (f) definition, however, "articles used for components of any such articles" as chewing gum and ordinary food poses a severe problem for the manufacturers' construction. Reading the parenthetical to exclude components of ordinary foods and chewing gum from the coverage of part (C) raises the anomalous possibility that a manufacturer might escape the premarketing approval requirements by distributing an otherwise regulable article not in pure form but as an ingredient of a food or chewing gum. See Nutrilab, 713 F.2d at 338 n. *. Such a result would convert a provision designed to bring a species of article not otherwise covered by any portion of the Act within the purview of the food provisions, see United States v. An Article of Food.... Aangamik 15, 678 F.2d 735 (7th Cir.1982); Natick Paperboard Corp. v. Weinberger, 525 F.2d 1103 (1st Cir.1975) cert. denied, 429 U.S. 819, 97 S.Ct. 65, 50 L.Ed.2d 80 (1976), into a boundless escape hatch from drug coverage. That reading would also conflict with the commonly understood import of the term "article" as used in the drug definitions, which the manufacturers recognize reaches any determinate substance, including one distributed as an ingredient of a nondrug product. Thus, contrary to the manufacturers' contention, the meaning of the term "food" for purposes of the parenthetical exclusion and its technical statutory meaning for purposes of the coverage of the food provisions cannot be identical. Cf. Nutrilab, 713 F.2d at 337-38 (noting that the statute "evidently uses the word `food' in two different ways," but wondering if exclusion refers "to `food' as a term of art in the statutory sense or to foods in their ordinary meaning?").
The government's construction, however, is equally problematic. The government in effect asks this court to impute the element of intention into the meaning of "food" in the parenthetical and thereby to confer a third meaning on the term distinct from those it carries in sections 321(f) and 321(f)(1). The language and structure of the relevant provisions seem inconsistent with this view. Neither the definition of food in subsection (f) nor the parenthetical exclusion of food in subsection (g)(1)(C) makes any reference to intention, see Nutrilab, 713 F.2d at 337, even though the concept is present in the definition in which the exclusion appears, 21 U.S.C. § 321(g)(1)(C); in another definition in the same subsection, id. § 321(g)(1)(B); and at other points in the section, id. § 321(h), (i), (s), (w), (x). The definition in section 321(g)(1)(B) instead refers to "articles used for food." By repeating the defined term in its own definition, and by making use or function the definitional *1505 criterion, Congress appears to have intended that this component of the statutory definition of "food" refer to common usage. The ordinary way in which an article is used, therefore, not any marketing claim on the part of the manufacturer or distributor as to a specific physiological purpose of that use, should determine whether it is a food for purposes of the parenthetical exclusion of section 321(g)(1)(C). See 1 H. Toulmin, The Law of Foods, Drugs and Cosmetics §§ 4.10, .15 (1963 & Supp.1969); cf. United States v. Technical Egg Products, Inc., 171 F.Supp. 326, 328 (N.D.Ga.1959) (item normally used for food is regulable as food under Act even though not intended for human consumption).
The government's suggested construction also renders the parenthetical in section 321(g)(1)(C) meaningless. If the subsection deems articles represented to affect the structure and functions of the human body drugs notwithstanding the common usage of those articles for food, then the coverage of an article by the food definition is irrelevant and the parenthetical excludes nothing that the nonparenthetical language would include. The government urges that, since all foods in some way affect bodily structure and function, the drafters must have inserted the parenthetical to avoid subjecting all foods to regulation as drugs by virtue of their universal effect on bodily structure and function. But the courts have always read the clauses in the statutory definitions employing the term "intended" to refer to specific marketing representations, see Action on Smoking and Health v. Harris, 655 F.2d 236, 238-39 (D.C.Cir.1980); National Nutritional Foods Association v. Mathews, 557 F.2d 325, 333-34 (2d Cir.1977), so this universal effect would not bring common foods sold without recourse to physiological claims within the statute in any case. As a logical matter, either the parenthetical limits the coverage of the subsection to nonfoods, or it expresses no limit beyond the nonparenthetical definition.
Construing the parenthetical to exclude all articles commonly used for food regardless of physiological claims made in connection with their sale does not defeat the central purpose of the provision as revealed in its legislative history. The original Food and Drug Act, enacted in 1906, classified as drugs only articles intended for therapeutic use. See Federal Food and Drug Act, ch. 3915, 34 Stat. 768 (1906). The legislative history of the 1938 Act makes clear that Congress added what became section 321(g)(1)(C) in order to expand the reach of the drug definition to cover products marketed for their physiological effects, which at the time escaped regulation. See 77 Cong.Rec. 5721 (1933) (statement of Senator Copeland upon introduction of S. 1944, 73 Cong., 1st Sess. (predecessor bill)) reprinted in 1 Legislative History of the Food, Drug, and Cosmetic Act 34; Foods, Drugs, and Cosmetics: Hearings on S. 2800 Before the Senate Committee on Commerce, 73d Cong., 2d Sess. 516 (1934) (statement of Walter Campbell, Chief of Food and Drug Administration, on subsequent predecessor bill), reprinted in 2 Legislative History 514, 519. Chief among these products in the drafters' minds were "slenderizers," see S.Rep. No. 493, 73d Cong., 2d Sess. 2 (1934), reprinted in 2 Legislative History 721, 722; 78 Cong.Rec. 8960 (1934) (statement of Senator Copeland), reprinted in 2 Legislative History 830, 831; Food, Drugs, and Cosmetics: Hearing Before a Subcomm. of the House Comm. on Commerce on H.R. 6906, H.R. 8805, H.R. 8941, and S. 5, 74th Cong., 1st Sess. 55 (1935) reprinted in 4 Legislative History 358, 370, particularly a nonfood preparation named "Marmola" which the Federal Trade Commission had unsuccessfully sought to regulate. See Hearings on S. 2800 at 516; Food, Drugs, and Cosmetics: Hearings on S. 1944 Before a Subcomm. of the Senate Comm. on Commerce, 73d Cong., 2d Sess. 15-16 (1933) (statement of Mr. Campbell), reprinted in 1 Legislative History 105, 108-08. See generally J. Young, The Medical Messiahs: A Social History of Health Quackery in Twentieth-Century America 122-28 (1967).
Much of the legislative history upon which the government relies, however, suggests that Congress acted not in order specifically *1506 to regulate as drugs all substances intended to affect bodily structure or function, but rather to reach those products, such as "Marmola," which evaded regulation altogether because they were neither foods nor therapeutic agents. See, e.g., United States v. 354 Bulk Cartons ... Trim Reducing-Aid Cigarettes, 178 F.Supp. 847 (D.N.J.1959). For example, the Chief of the Food and Drug Administration testified during hearings on a predecessor bill that
[t]here are products on the market now that escape control either under the definition of "food" or under the definition of "drugs" in the present act, such as slenderizing products, reducing products. Obesity is not itself a disease in all instances and products advocated and sold for the treatment of obesity, as a matter of fact, are not always subject to the terms of this act.
Hearings on S. 2800 at 519. Other pieces of legislative history cited by the government simply track the statutory language. They establish that Congress meant to regulate drugs intended to produce a physiological effect, but do not bear on the distinction between food and drug which lies at the heart of this dispute. For example, the House committee report on the bill finally enacted stated:
Drugs intended for ... remedying underweight or overweight or for otherwise affecting bodily structure or function are subjected to regulation ....
The definition of drug is expanded to include ... articles other than food intended to affect the structure or any function of the body of man or other animals. [This expansion is] needed to give jurisdiction over a great number of drugs which are not amenable to control under the present law.
H.R.Rep. No. 2139, 75th Cong., 3d Sess. 2, 3 (1938), reprinted in 6 Legislative History 300, 301, 302.
In addition, the legislative history indicates that "slenderizing foods" were among the substances which Congress sought to reach under the provisions of the Act specifying labeling requirements for food which "purports to be or is represented for special dietary use." See 21 U.S.C. § 343(j); S.Rep. No. 361, 74th Cong., 1st Sess. 12 (1935), reprinted in 3 Legislative History 660, 671. The contemporaneous addition of this provision no more demonstrates that all weight-reduction products are foods, as the manufacturers here originally urged, than do the references to drugs intended to affect structure or function establish that all substances intended to have such effects are drugs. But adding the provision does suggest that Congress contemplated different treatment for articles intended as aids in achieving the physiological effect of weight reduction, depending upon their status as foods or drugs.
The piece of legislative history most strongly supporting the government's interpretation appears in the report of the Commerce Committee which accompanied S. 2800, one of the early predecessor bills. Discussing the expanded definition of drug, the report states:
[T]he provision contained in that draft that the definition of food, drug, and cosmetic should not be construed as mutually exclusive, has been deleted for the reason that this language is superfluous. The present law does not have such a clause relating to the definitions of food and drug and there has never been a court decision to the effect that these definitions are mutually exclusive, despite the fact that repeated actions have been brought, for example, against filthy foods bearing unwarranted therapeutic claims, alleging these products to be adulterated as a food because of their filth and misbranded as drugs because of their false and fraudulent therapeutic claims.
The use to which the product is to be put will determine the category into which it will fall. If it is to be used only as a food it will come within the definition of food and none other. If it contains nutritive ingredients but is sold for drug use only, as clearly shown by the labeling and advertising, it will come within the definition of drug, but not that of food. If it is sold to be used both as a food and *1507 for the prevention or treatment of disease it would satisfy both definitions and be subject to the substantive requirements for both. The manufacturer of the article, through his representations in connection with its sale, can determine the use to which the article is to be put. For example, the manufacturer of a laxative which is a medicated candy or chewing gum can bring his product within the definition of drug and escape that of food by representing the article fairly and unequivocally as a drug product.
S.Rep. No. 493, 73d Cong., 2d Sess. 2-3 (1934), reprinted in 2 Legislative History, 721, 722-23. This discussion does not necessarily contemplate dual classification under section 321(g)(1)(C), however. The report notes that the drafters deleted as "superfluous" a clause which provided that the various definitions not be construed as mutually exclusive, but the previous experience under the Act which prompted them to do so related only to the therapeutic definition. Indeed, the entire discussion mentions only foods "bearing unwarranted therapeutic claims" and those "to be used both as a food and for the prevention of or treatment of disease;" it does not explicitly discuss either the parenthetical or the possibility of a common food marketed for its physiological effect. And the specific example of "a laxative which is a medicated candy or chewing gum" is inconclusive, given that a "medicated" laxative would be regulable as a drug because of the presence of a determinate ingredient independently so classified.
Finally, a court's responsibility to construe the statute in accord with its protective purposes does not confer a license to ignore congressional judgments reflected in the classification scheme. See 62 Cases of Jam v. United States, 340 U.S. 593, 600, 71 S.Ct. 515, 520, 95 L.Ed. 566 (1951); NNFA v. Mathews, 557 F.2d at 336-37. Items classified as foods by no means escape regulation. Though food manufacturers need not obtain premarketing approval for their products, they are still subject to the Act's provisions on adulteration and misbranding. 21 U.S.C. §§ 342, 343. To enforce these provisions, the FDA may inspect factories, id. § 374; commence seizures of adulterated foods, id. § 334; seek injunctions against the sale of adulterated or misbranded food, id. § 332; and seek criminal penalties in appropriate cases, id. § 333. Though these provisions do not bear directly on the threshold question whether an item is a food or a drug, they do support the inference that Congress determined that a different level of regulation was adequate to protect the public in the case of an article commonly used for food, even though marketers of the product claim that it produces specific physiological effects. Absent specific indications undermining that inference in the legislative history, such as those unearthed by the AMP and Bacto-Unidisk courts with respect to "devices," the congressional determination deserves respect.
Thus, if an article affects bodily structure or function by way of its consumption as a food, the parenthetical precludes its regulation as a drug notwithstanding a manufacturer's representations as to physiological effect. The Act evidences throughout an objective to guarantee accurate information to consumers of foods, drugs, and cosmetics. See, e.g., 21 U.S.C. § 343. The presence of the parenthetical in part (C) suggests that Congress did not want to inhibit the dissemination of useful information concerning a food's physiological properties by subjecting foods to drug regulation on the basis of representations in this regard.
Whether or not the parenthetical of section 321(g)(1)(C) incorporates the entire technical definition of food in section 321(f), however, and whether or not section 321(g)(1)(C) contemplates dual classification, the manufacturers cannot prevail here. Notwithstanding the government's contention that the FDA could regulate their products as a food if it chose to do so, the manufacturers cannot demonstrate that starchblockers are a food in any sense cognizable under the statute.
II. Applying Section 321(g)(1)(C) to Starchblockers.
The Seventh Circuit recently considered the identical question of the status of *1508 starchblockers under part (C) and concluded that they are drugs. Nutrilab, Inc. v. Schweiker, 713 F.2d 335 (7th Cir.1983), aff'g, 547 F.Supp. 880 (N.D.Ill.1982). The Nutrilab district court had treated the issue as one of intended use, and held that "`food' refers only to those items actually and solely ... consumed either for taste, aroma, or nutritional value." 547 F.Supp. at 883. The Circuit Court found this definition "unduly restrictive" because, it observed, "some products such as coffee or prune juice are undoubtedly food but may be consumed on occasion for reasons other than taste, aroma, or nutritive value." 713 F.2d at 338. It instead defined food as articles used "primarily for taste, aroma, or nutritive value," id. (emphasis added), properly rejecting also any suggestion that the source of a product makes it a food, id. at 337.
Here the manufacturers contend that starchblockers must be deemed a food because their biochemical composition varies from that of the bean flour used for making bread a paradigmatic food only by the percentage of each component and the addition of excipients and binders. This argument fails for the same reasons articulated in Nutrilab. The concentration of certain components during processing effects a significant physical change. The Supreme Court recently ruled in fact that the marketing of an established drug with different excipients and binders will necessitate submission of an application for approval as a new drug. See United States v. Generix Drug Corp., ___ U.S. ___, 103 S.Ct. 1298, 75 L.Ed.2d 198 (1983). Most fundamentally, the argument fails to address the Act's focus on usage.
The government approves the Seventh Circuit's result in Nutrilab, but takes issue with the standard of usage by which it was reached. The government argues that the meaning of "food" for purposes of the parenthetical exclusion of section 321(g)(1)(C) diverges from the definition of that term contained in section 321(f). If a type of article is used at all in a manner associated with the ordinary meaning of food for taste, aroma, or nutritive value the government states it may be regulated as a food under section 321(f)(1). On the other hand, if the same food were represented to have a particular structural or functional effect on the body, the government claims it may also be regulated as a drug under section 321(g)(1)(C). Thus, the government prefers the formulation of the Nutrilab district court, which held that only if a food is used solely for taste, aroma, or nutrition and no claims are made that it affects bodily function may it escape regulation under section 321(g)(1)(C).
The manufacturers concur in the Seventh Circuit's criticism of the Nutrilab district court's definition, but contend that the appellate court's formulation is itself "unduly restrictive" because it fails to draw appropriate conclusions from the fact it recognizedthat some foods, such as coffee and prune juice, are frequently consumed specifically for their physiological effect, but are nonetheless foods. But this criticism, even if it were valid, has no relevance to the present case. The manufacturers adduced no evidence at trial that starchblockers are ever consumed by anyone for taste, aroma, nutrition, or sustenance. They manufacture their products by a process which concentrates the antinutrient to the exclusion of components which contribute food value. One could adjust the definitional threshold for "food" considerably lower than the Seventh Circuit's "primarily used" standard and still exclude starchblockers, because they have no taste or aroma, their nutritional value is negligible, they provide no sustenance, and they are not consumed for any of these purposes.
An example offered by the manufacturers illustrates well the consequences of the dual classification dispute, but demonstrates at the same time its irrelevance to the case of starchblocker pills. Coffee is often consumed as a stimulant, but is also commonly drunk for its taste and aroma, if not its nutritional value. Were the parenthetical read to preclude dual classification, the common usage of coffee as a food would place it beyond the reach of the part (C) drug definition even though a manufacturer might promote its coffee exclusively for its use in staying awake. (A doctor *1509 who testified at the hearing on behalf of the FDA stated that such a coffee would remain a food notwithstanding such promotional claims.) This result is unavailable to the manufacturers of starchblockers, however, because they cannot demonstrate that their products are commonly used for food. The manufacturers simply cannot bring their products within the parenthetical of part (C).
The manufacturers also attempt to draw an analogy between starchblockers and saccharin, suggesting that each, though without food value itself, makes other foods more palatable to the dieter. See 21 C.F.R. § 105.3(a)(2) (artificial sweeteners as food for special dietary use). The analogy fails because saccharin is commonly consumed for its taste effect. Though the taste of saccharin may have food value only when used to enhance the flavor of other foods, it is a food characteristic nonetheless. Garlic is no less enticing because people generally do not eat it by the clove.
Finally, the manufacturers contend that "demonstrated uselessness" as a food does not render a product a drug, citing as support the decision in National Nutritional Foods Association v. Mathews, 504 F.2d 761, 786-89 (2d Cir.1974). In NNFA v. Mathews, the Second Circuit rejected the FDA's attempt to regulate as drugs under section 321(g)(1)(B) all vitamin and mineral products in excess of the upper limits of the U.S. Recommended Daily Allowances. Its analysis does not bolster the manufacturers' position, however. First, because part (B) clearly countenances dual classification, the case did not involve the status of the products as food, and the inquiry into nutritional value bore only on the question of therapeutic intent. Here, by contrast, the manufacturers concede the requisite intent. Second, the manufacturers quote the Circuit Court's language out of context. The court stated that "[w]hile ... a factfinder should be free to pierce all of a manufacturer's subjective claims of intent and even his misleadingly `nutritional' labels to find actual therapeutic intent on the basis of objective evidence in a proper case, such objective evidence would need to consist of something more than demonstrated uselessness as a food for most people." Id. at 789 (emphasis added). The court had specifically found that "a significant number of persons have indisputable nutritional need for potencies exceeding the upper limits," id., which established that these products were "used for food" to at least some extent. Starchblockers, by contrast, are utterly useless for any food purpose, which renders meritless their manufacturers' contention that the FDA should classify them as food. In short, starchblockers are a drug under section 321(g)(1)(C), whatever may be the precise breadth of the parenthetical exclusion of that part.
Plaintiff Sunrise Chemical, Inc., does not make starchblocker tablets but rather processes the flour described variously as a "legume, protein concentrated food substance," Complaint ¶ 12, and a "high protein bean flour," Sklar Affidavit in Opposition to Motion for Preliminary Injunction ¶ 2 from which they are made. At the hearing counsel for the manufacturers suggested that Sunrise might have a different status under section 321(g)(1)(C) than makers of the pills. Counsel presented little evidence to support this contention, however. The company "applaud[s] and approve[s] of the use of its product by others in terms of promotion for weight control regimens," id. ¶ 7, so it cannot contend that its product is not "intended to affect the structure or any function of the body of man." Further, though the air-classified, protein-concentrated flour which Sunrise purchases from its supplier can be used to make food products such as bread, see id.; Transcript at 276-90, Sunrise concedes that it further processes the flour in a manner which makes it "more suitable for free-flow tableting, encapsulating or other form [sic]," id. ¶ 5. A bean flour suitable for use in the manufacture of food might be immune to regulation under section 321(g)(1)(C). But the sparse evidence adduced at the hearing suggests that Sunrise's product is specifically and perhaps exclusively prepared for use as starchblockers, and is therefore a proper subject for seizure or other regulation.
*1510 For the foregoing reasons, starchblocker pills are declared drugs under the Act. Their seizure is therefore permissible. The Clerk will enter judgment accordingly, without costs, in this case and in all presently pending, related cases. Any party that manufacturers not starchblockers but a bean flour actually used for foods may move for appropriate relief upon a proper showing.
SO ORDERED.
NOTES
[1] The following is a list of all related cases pending before this Court:
United States v. An Article of Drug, No. 82 Civ. 6290 (S.D.N.Y. filed Aug. 21, 1982);
United States v. An Article of Drug, No. 83 Civ. 422 (S.D.N.Y. filed Jan. 1, 1983, upon transfer from N.D.Ill.);
United States v. Articles of Drug, No. 83 Civ. 1819 (S.D.N.Y. filed Mar. 9, 1983, upon transfer from W.D.Wash.);
United States v. Article of Drug, No. 83 Civ. 1940 (S.D.N.Y. filed Mar. 14, 1983, upon transfer from W.D.Pa.);
United States v. Advantage Starch Blocker, No. 83 Civ. 2392 (S.D.N.Y. filed Mar. 28, 1983, upon transfer from E.D.N.C.);
United States v. An Article of Drug, No. 83 Civ. 2537 (S.D.N.Y. filed Apr. 4, 1983, upon transfer from N.D.Okl.);
United States v. Article of Drug, No. 83 Civ. 2742 (S.D.N.Y. filed Apr. 11, 1983, upon transfer from S.D.Ohio);
United States v. Article of Drug, No. 83 Civ. 2743 (S.D.N.Y. filed April 11, 1983, upon transfer from N.D.Tex.);
United States v. An Article of Drug, No. 83 Civ. 2746 (S.D.N.Y. filed April 11, 1983, upon transfer from S.D.Ohio);
United States v. An Article of Drug, No. 83 Civ. 2879 (S.D.N.Y. filed Apr. 14, 1983, upon transfer from N.D.Tex.);
United States v. An Article of Drug, No. 83 Civ. 2901 (S.D.N.Y. filed Apr. 15, 1983, upon transfer from W.D.Pa.);
United States v. An Article of Drug, No. 83 Civ. 3054 (S.D.N.Y. filed Apr. 20, 1983, upon transfer from W.D.N.Y.);
United States v. An Article of Drug, No. 83 Civ. 3185 (S.D.N.Y. filed Apr. 26, 1983, upon transfer from N.D.Tex.);
United States v. An Article of Drug, No. 83 Civ. 3314 (S.D.N.Y. filed Apr. 29, 1983, upon transfer from N.D.N.Y.);
United States v. An Article of Drug, No. 83 Civ. 3757 (S.D.N.Y. filed May 17, 1983, upon transfer from N.D.Tex.);
United States v. An Article of Drug, No. 83 Civ. 3977 (S.D.N.Y. filed May 23, 1983, upon transfer from E.D.Tenn.);
United States v. An Article of Drug, No. 83 Civ. 4265 (S.D.N.Y. filed June 6, 1983, upon transfer from S.D.Fla.);
United States v. An Article of Drug, No. 83 Civ. 4346 (S.D.N.Y. filed June 9, 1983, upon transfer from S.D.Ill.);
United States v. An Article of Drug, No. 83 Civ. 4595 (S.D.N.Y. filed June 20, 1983, upon transfer from N.D.Ga.);
United States v. An Article of Drug, No. 83 Civ. 5122 (S.D.N.Y. filed July 11, 1983, upon transfer from W.D.N.C.);
United States v. An Article of Drug, No. 83 Civ. 6186 (S.D.N.Y. filed Aug. 18, 1983, upon transfer from D.N.J.);
United States v. An Article of Drug, No. 83 Civ. 6422 (S.D.N.Y. filed Aug. 30, 1983, upon transfer from D.Utah);
United States v. An Article of Drug, No. 83 Civ. 6759 (S.D.N.Y. filed Sept. 14, 1983, upon transfer from D.Conn.);
United States v. An Article of Drug, No. 83 Civ. 6965 (S.D.N.Y. filed Sept. 23, 1983, upon transfer from D.N.J.);
United States v. An Article of Drug, No. 83 Civ. 7123 (S.D.N.Y. filed Sept. 28, 1983, upon transfer from D.N.J.).
Further transfers of cases to this District will be inappropriate.
[2] Questions of safety and effectiveness of the manufacturers' products do not bear directly on this case at the present stage. See, e.g., United States v. An Article of Drug .... Bacto-Unidisk, 394 U.S. 784, 791-92, 89 S.Ct. 1410, 1414-15, 22 L.Ed.2d 726 (1969); United States v. Diapulse Corp. of America, 457 F.2d 25, 27-28 (2d Cir. 1972). However, a manufacturer may avoid the burden of submitting a new drug application if its product is generally recognized as safe and effective, see 21 U.S.C. § 321(p), and a district court may exercise jurisdiction to decide this issue. Premo Pharmaceutical Laboratories, Inc. v. United States, 629 F.2d 795, 801 (2d Cir.1980).
The evidence presented demonstrated that no consensus exists. The FDA presented evidence of seventy-five reports of adverse reactions to the ingestion of starchblockers. See Troendle Affidavit; Nutrilab, Inc. v. Schweiker, 713 F.2d 335, 336 (7th Cir.1983). Many of these, however, were telephone calls relating complaints of gastrointestinal upset, coinciding with the agency's announcement in July 1982 that it would henceforth treat starchblockers as drugs. The FDA's own experts were unable to conclude from these reports that starchblockers posed a significant health hazard. See Transcript at 183-85; Exhibit D at 3-4. Considerable evidence also raised the possibility of long-term adverse health effects from starchblockers. Compare Transcript at 36-37, 55-56, 83 (Dr. Callaway) (starchblockers might induce enlargement of pancreas and pancreatitis as indirect result of blocking trypsin); id. at 150-51 (Dr. Morrow); id. at 16, 19 (Dr. Liener) (hemagglutinins may damage cells lining intestinal tract), with id. at 240-42 (only hemagglutinins in red beans produce toxic effects); id. at 225-32 (trypsin inhibitors only weakly inhibit human trypsin and probably have no effect on human pancreas); id. at 294-95 (common foods, such as string beans, bagels, and rye bread, contain many times more hemagglutinins than starchblockers). The manufacturers called two experts who testified that they had conducted studies on animals which produced no evidence of toxicity. Id. at 318-29 (Dr. Gibson), 349-58 (Dr. Wallin).
The evidence as to efficacy seems equally contradictory. One published study concluded that "starch-blocker tablets do not inhibit the digestion and absorption of starch calories in human beings." Bo-Linn, Santa Ana, Morawski & Fordtran, Starch Bockers Their Effect on Calorie Absorption From a High-Starch Meal, 307 New Eng.J.Med. 1413, 1413 (1982). Another concluded that pure alpha-amylase inhibitor "did not affect the hydrolysis of dietary starch as measured by the ability of dietary starch to provide energy for the growing rat." Savaiano, Powers, Costello, Whitaker & Clifford, The Effect of an Alpha-Amylase Inhibitor on the Growth Rate of Weanling Rats, 15 Nutrition Rep. Int'l 443, 443 (1977); see Transcript at 100-03, 126-27 (Dr. Whitaker). The manufacturers, on the other hand, presented evidence of an unpublished study with dogs which determined that "starch digestion was partially inhibited" by contemporaneous consumption of the "legume protein concentrate product" of plaintiff AMX. See Affidavit of Rene Gibson ¶ 7. Dr. Callaway testified, however, that neither blood glucose nor fecal carbohydrate levels, the bases of Dr. Gibson's conclusions, were reliable indicators of starch digestion. See Transcript at 52-54; see also id. at 311 (Dr. Thompson) (blood glucose level more reliable measure than fecal calorie measure). The manufacturers also presented evidence of a human study conducted by plaintiff General Nutrition Corp. that indicated that "consumption of its product ... with a starch meal ... results in a reduction of as much as 22% in the digestion of calories." Thompson Affidavit ¶ 7. FDA experts criticized this study also for its reliance on blood glucose levels. See Transcript at 48, 53, 60-61 (Dr. Callaway); 398-99 (Dr. Bo-Linn); see also id. at 311 (Dr. Thompson). Thus, the manufacturers have by no means established that a consensus exists as to the effectiveness of starchblockers for their claimed purpose.