MANSFIELD, Circuit Judge:

In this action by Premo Pharmaceutical Laboratories Inc. (Premo), a manufacturer of drugs, for a declaratory judgment and injunctive relief, 28 U.S.C. §§ 2201, 1331(a), the Government, the Secretary of Health, Education and Welfare and the Commissioner of Food and Drugs appeal from a decision and judgment of the District Court for the Southern District of New York entered on August 2, 1979, by Judge Milton Pollack and reported at 475 F.Supp. 52, declaring (1) that "Insulase," a drug product manufactured by plaintiff-appellee, Premo, for use in the treatment of diabetes, is not a "new drug" within the meaning of § 201(p) of the Food, Drug and Cosmetic Act, 21 U.S.C. § 321(p), (2) that any regulatory action by the Government on the basis that Insulase is a "new drug" is unlawful, and holding that Premo may market Insulase without obtaining approval of its new drug application from the Food and Drug Administration (FDA), filed pursuant to 21 U.S.C. § 355(a). We reverse and remand with directions to dismiss the complaint.

The active ingredient in Premo's Insulase is chlorpropamide (CPA), which is also the active ingredient in "Diabinese," an FDA-approved drug product which has been made and marketed by Pfizer Laboratories, Inc. for many years. However, the inactive ingredients in the two drugs, known as excipients, differ. Excipients are typically added to an active ingredient to form a tablet, capsule coating, coloring or flavor. As the district court found, although an inactive ingredient may by itself be safe, it may, when combined with an active drug ingredient, affect a drug's safety and effectiveness.

Under § 505 of the Act, 21 U.S.C. § 355, no person may market a new drug unless he files with the FDA a new drug application (NDA) demonstrating that the drug is both safe and effective for the use for which it is intended and obtains FDA approval. Normally the applicant furnishes controlled chemical tests and investigations showing that the product is safe and effective, 21 U.S.C. § 355(d). But where the drug product is claimed to be a copy of one already approved by the FDA on the basis of such submissions — sometimes called a "me-too" drug — the applicant may file with the FDA an "abbreviated new drug application" (ANDA), which relies upon the safety and effectiveness tests conducted with respect to the FDA-approved drug (sometimes called the "pioneer drug"). The FDA will only approve an ANDA, however, where the "me-too" drug product is shown to be the therapeutic equivalent of the pioneer and safe and effective in accordance with 21 U.S.C. § 355(d). See generally, Hoffman-LaRoche, Inc. v. Weinberger, 425 F.Supp. 890 (D.D.C. 1975).

In the present case Premo in 1978 first filed with the FDA an ANDA for Insulase, submitting among other things a comparative bioavailability study with respect to the Insulase and Pfizer's Diabinese. The FDA found the data insufficient to establish therapeutic bioequivalence and requested further evidence demonstrating the performance of Insulase under conditions similar to actual use. Premo chose not to comply and began marketing Insulase without FDA approval, which led the FDA in December, 1978, to file seizure actions against Insulase in five United States district courts pursuant to 21 U.S.C. § 334 on the ground that Insulase was a new drug being marketed without prior FDA approval.

In the meantime, in November, 1978, Premo had commenced the present action seeking to enjoin the FDA from instituting any regulatory action against it under 21 U.S.C. § 334 and in January, 1979 it amended its complaint to seek a declaratory judgment to the effect that Insurance was not a "new drug" within the meaning of 21 U.S.C. § 321(p). It also sought an injunction restraining the defendants from instituting any seizure action under the Act. The court denied preliminary injunction on the ground that there were unresolved questions as to the safety and effectiveness of Insulase and held a trial to determine whether Insulase was a "new drug."

At trial Premo contended that the term "drug" as used in "new drug" refers only to the active ingredient in the product, as distinguished from excipients. Judge Pollack properly rejected this argument, pointing out that "drug" as defined in § 201(g)(1)(B) of the Act, 21 U.S.C. § 321(g)(1)(B) encompasses drug products as well as active ingredients, 475 F.Supp. at 54, and that "differences in excipients may impair the safety or effectiveness of a drug product even though its active ingredient is generally recognized as safe and effective." Id. at 55. See in accord Pharmadyne Laboratories, Inc. v. Kennedy, 466 F.Supp. 100, 104 (D.N.J.), aff'd., 596 F.2d 568, 571 n.6 (3d Cir. 1979). The Government contended, relying on the language of 21 U.S.C. § 321(p), that any drug product constitutes a "new drug" unless it is shown that the drug is "generally recognized, among experts qualified by scientific training and experience to evaluate the safety and effectiveness of drugs, as safe and effective for use . . ." This reliance on the plain language of the statute was rejected by the district court on the ground that it would require FDA approval of all drug products, frustrating the "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 55. The district court ruled that where the active ingredient in a questioned product is the same as that in a product already approved by the FDA as safe and effective, but the excipients differ, the questioned product is not a "new drug" if the excipients are "generally recognized individually to be safe" and "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. at 55.

The district court's approach to the "new drug" issue resulted in its undertaking the task of resolving complex scientific issues as to whether Insulase had the same bioavailability as Diabinese and thus was a safe and effective therapeutic equivalent, rather than confining itself to determining whether, based on material usage over a substantial period of time, and published scientific studies or other publicly available data, scientific experts "generally recognized" Insulase to be safe and effective and therefore not a "new drug." The trial was reduced largely to a battle of scientific experts centered around private, unpublished tests made by Premo to determine the bioavailability of Insulase as compared with Diabinese. Premo's experts were of the opinion, based on these studies, that subject to some qualifications the two drugs were therapeutically equivalent and therefore interchangeable. The Government experts criticized severely the analytical methods employed by Premo's researchers in making the studies, the use by Premo of a computer to predict steady blood levels of Insulase rather than of a multiple dosage test, and the failure to use a "titration" procedure to determine the amount of drug in a patient's blood. They also differed in their interpretation of the results of the studies. Since the studies had not been published and most (if not all) of the witnesses had learned of Insulase only in preparation for trial, none of the experts, including Premo's witnesses, could testify that Insulase was generally recognized as safe and effective or as the therapeutic equivalent of Diabinese.

Judge Pollack, using his own definition of a "new drug," found that Premo's bioavailability study "was conducted in accordance with accepted and reliable scientific methods," 475 F.Supp. at 56, that it was "of no clinical significance that Diabinese delivers more CPA [chlorpropamide] than Insulase after the first tablet is taken," that the "differences in the steady state blood levels achieved by the two products are not significant," and "that Insulase is not less safe and effective than Diabinese." Based on these scientific findings he concluded that the two drugs were therapeutic equivalents and that Insulase was therefore not a "new drug." The district court did not withdraw or change its earlier finding made in denying a preliminary injunction to the effect that

"[T]he specific combination of active drug and excipients which constitutes Insulase are not at this time 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." (A. 301).

Accordingly the district court entered a judgment declaring that Premo's Insulase "is not a new drug within the meaning of the Federal Food, Drug and Cosmetic Act" and that "Regulatory action by the defendants against Premo and/or Premo's chlorpropamide's [sic] tablets on the basis that said product is a new drug is unlawful." From this judgment defendants appeal.

DISCUSSION

At the outset we must consider whether the district court had subject matter jurisdiction over this lawsuit, in view of the pendency of seizure actions instituted against Insulase by the Government in five other districts and the pendency of Premo's ANDA before the FDA for approval of Insulase as safe and effective. Had the district court issued an order enjoining the multiple seizure cases, we might well dismiss the present action in accordance with the well settled principle that federal courts lack jurisdiction to enjoin seizure actions instituted by the FDA under 21 U.S.C. § 334. Ewing v. Mytinger Casselberry, Inc., 339 U.S. 594, 601, 70 S.Ct. 870, 873, 94 L.Ed.2d 1088 (1950). See also Abbott Laboratories v. Gardner 387 U.S. 136, 146-48, 87 S.Ct. 1507, 1514-15, 18 L.Ed.2d 681 (1967); Pharmadyne Laboratories, Inc. v. Kennedy, 596 F.2d 568 (3d Cir. 1979). Here, however, the district court denied injunctive relief and instead merely issued a declaratory judgment. Declaratory judgment actions on "new drug" issues have been entertained by the courts in the past, see A. M. P. 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 the applicability of Ewing to declaratory judgments — as opposed to injunctive relief — is unsettled. The rationale underlying Ewing is that the Government should not be prematurely stopped from seizing potentially dangerous drugs as long as the aggrieved party can litigate the merits in an enforcement proceeding. Here, in contrast, the Government remained free to seize plaintiff's product until such time as the merits were litigated.

We turn to Pfizer's contention as an amicus curiae that the present action should have been dismissed because of Premo's failure to exhaust its administrative remedies before the FDA. Pfizer is not really asserting exhaustion in the traditional sense of an administrative remedy to be pursued prior to proceeding in the district court. The import of Pfizer's contention is that the merits of whether an NDA is required should have been determined by the FDA because the district court either lacked jurisdiction or should not have exercised its jurisdiction as a matter of discretion. Although the FDA has jurisdiction to decide the issue before us, Weinberger v. Hynson, Westcott Dunning, Inc., 412 U.S. 609, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973); Weinberger v. Bentex Pharmaceuticals, Inc., 412 U.S. 645, 93 S.Ct. 2488, 37 L.Ed.2d 235 (1973), and the district court may refer a "new drug" issue to the FDA for resolution, either in its consideration of Premo's pending ANDA proceeding or upon separate petition by Premo to the FDA pursuant to 21 C.F.R. § 10.30, Weinberger v. Bentex Pharmaceuticals, Inc., supra, 412 U.S. at 653-54, 93 S.Ct. at 2494 ("the District Court's referral of the `new drug' . . issues to FDA was appropriate, as these are the kind of issues peculiarly suited to initial determination by the FDA."), the FDA's jurisdiction is not exclusive, see Ciba Corp. v. Weinberger, 412 U.S. 640, 644, 93 S.Ct. 2495, 2498, 37 L.Ed.2d 230 (1973), in appropriate circumstances a court may exercise its concurrent jurisdiction to adjudicate the "new drug" status of a product. In this case we conclude that the limited issue on the merits requiring decision was sufficiently clear to warrant the district court's exercise of its subject matter jurisdiction. Moreover, to refer the issue to the FDA at this late date would be wasteful and duplicative.

Turning to the merits, under the express language of 21 U.S.C. § 321(p), which defines the term "new drug" as used in the Act, it is clear that, with some exceptions not pertinent here, a drug product must be classified as a "new drug" unless its manufacturer can show (1) that it is "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" and (2) that being so recognized it has "been used to a material extent or for a material time under such conditions." Hynson, supra, 412 U.S. at 631, 93 S.Ct. at 2484 (". . . a drug cannot transcend `new drug' status until it has been used `to a material extent or for a material time.'"). Unless the pharmaceutical manufacturer can show that its drug product meets these conditions it must file with the FDA a new drug application (either an NDA or an ANDA, depending on whether the drug is a "pioneer" or a "me-too" product) and establish by substantial evidence to the FDA's satisfaction that the drug is safe and effective for its intended uses, Hynson, supra, 412 U.S. at 629-30, 93 S.Ct. at 2483. Should the FDA deny approval, the applicant is entitled to an administrative hearing and judicial review on the basis of the administrative record, 21 U.S.C. § 355(h). The purpose of the Act, as enacted in 1938 and amended by Congress in 1962, is to protect the public against danger to human life arising from use of unsafe and ineffective drugs by assuring that before any drug is marketed it will have been carefully reviewed by FDA experts. Congress' exclusion of "generally recognized" drug products from the definition of a "new drug" is a very narrow one, which is not intended to permit a pharmaceutical manufacturer to substitute its opinion regarding the safety or effectiveness of a drug for that of the FDA, the publicly recognized repository of expertise in such matters, or to require the court to develop its own body of scientific knowledge in substitution for that of the FDA. Indeed, the history of the exemption confirms its limited scope.

Prior to the enactment of the 1938 Act there was no requirement that a drug be approved by the FDA before being marketed. As a result, there were in 1938 some unsafe drugs on the market and some which had been used for so many years and had been so well tested that they were generally recognized by experts qualified in drug evaluation to be safe. Congress therefore exempted these drugs from the definition of "new drugs" requiring FDA approval. All other drugs, including those as to which there was any dispute among experts about their safety, would of course be new drugs and could be sold only after FDA review and approval.

When Congress in 1962 amended the Act to require that a drug be shown to be effective as well as safe, the same situation existed with respect to the effectiveness of drug products being marketed in 1962 as had existed regarding the safety of drugs being sold at the time of the enactment of the Act in 1938. The exemption therefore embraced only drugs that were generally recognized by qualified experts to be both safe and effective. Upon enactment of the 1962 amendments the FDA arranged with the National Academy of Sciences/National Research Council to create panels to review, study and establish the efficacy of all outstanding drug products, whether or not previously approved, based on the best available data, Hynson, supra, 412 U.S. at 614-15, 93 S.Ct. at 2475-76. When the panels found that a drug product was effective for certain uses, the FDA publicly announced in a Drug Efficacy Study Implementation Notice those uses for which the drug could be approved without the necessity of an NDA.

For present purposes, the important point is that only those pre-existing drug products found through careful study by scientific experts to have been generally recognized on the basis of usage and documentation to be safe and effective were entitled to the exemption. The mere fact that a pharmaceutical manufacturer purports to market a "me-too" product does not exclude the drug from the definition of "new drug." Otherwise, the manufacturer could, based on its own opinion founded on its studies of its own so-called "me-too" product, market a drug product that has not gained such general recognition. Moreover, either the unawareness of the drug product by experts generally or a genuine dispute among qualified experts regarding a drug product's safety and effectiveness preclude its qualifying for exclusion as "generally recognized." United States v. An Article of Drug — Furestrol, 294 F.Supp. 1307, 1311 (N.D.Ga. 1968), affd. on the opinion below, 415 F.2d 390, 392 (5th Cir. 1969); United States v. An Article of Drug — Mykocert, 345 F.Supp. 571, 575 (N.D.Ill. 1972); see United States v. An Article of Drug — X-Otag Plus Tablets, 441 F.Supp. 105, 110 (D.Col. 1977), affd. in part and remanded in part, 602 F.2d 1387 (10th Cir. 1979); United States v. Seven Cartons, 293 F.Supp. 660, 662-63 (S.D.Ill. 1968), affd. in part and vacated in part, 424 F.2d 1364 (7th Cir. 1970). As the Supreme Court stated in Hynson, supra, 412 U.S. at 632, 93 S.Ct. at 2484:

"We accordingly have concluded that a drug can be `generally recognized' by experts as effective for intended use within the meaning of the Act only when that expert consensus is founded upon `substantial evidence' as defined in [ 21 U.S.C. § 355(d)]."

Normally the general expert consensus is supported not only by clinical experience but by publication in the scientific literature. Bentex, supra, 412 U.S. at 652, 93 S.Ct. at 2493. "[T]he Act is designed so that drugs on the market . . . will have mustered the requisite scientifically reliable evidence of effectiveness long before they are in a position to drop out of active regulation by ceasing to be a `new drug'." Hynson, supra, 412 U.S. at 631, 93 S.Ct. at 2484. As the Supreme Court stated in Bentex, supra, 412 U.S. at 652, 93 S.Ct. at 2493:

"Whether a particular drug is a `new drug,' depends in part on the expert knowledge and experience of scientists based on controlled clinical experimentation and backed by substantial support in scientific literature. One function is not peculiar to judicial expertise, the other to administrative expertise. The two types of cases overlap and strongly suggest that Congress desired that the administrative agency make both kinds of determination. Even where no such administrative determination has been made and the issue arises in a district court in enforcement proceedings, it would be commonplace for the court to await an appropriate administrative declaration before it acted."

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 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. Where an unapproved product not believed to be excluded from "new drug" classification is being marketed, the FDA is vested with the power to institute seizure actions. Nothing in the language of the Act or its legislative history 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. As we stated in AMP, Inc. v. Gardner, supra, 389 F.2d at 831:

"But the safety of the products is not what is at issue here. The question is whether there is general recognition among qualified experts of the products' safety and effectiveness — if there is not, the products must be submitted to the Secretary of Health, Education and Welfare for a determination as to the safety, adequacy for testing, etc."

Accord, United States v. 1,048,000 Capsules — Afrodex, 494 F.2d 1158, 1160 (5th Cir. 1974); Tyler Pharmacal Distributors, Inc. v. U.S. Department of Health, Education and Welfare, 408 F.2d 95, 99 (7th Cir. 1969); United States v. Articles of Drug — Colchicine, 442 F.Supp. 1236, 1242 (S.D.N.Y. 1978), affd. mem. sub nom. United States v. Articles of Drug — 100 Dihycon (2d Cir., April 6, 1979), Dkt. No. 78-6071.

"Under the scheme of the Act the ultimate determination of the safety of a drug is not a matter given to the courts, but one to be determined by the Food Drug Administration upon submission of an NDA." United States v. 1,048,000 Capsules — Afrodex, supra, 494 F.2d at 1160.

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. See Far East Conference v. United States, 342 U.S. 570, 574, 72 S.Ct. 492, 494, 96 L.Ed. 576 (1952). As the Court stated in Bentex, supra, 412 U.S. at 653-54, 93 S.Ct. at 2494:

"As the District Court said: `Evaluation of conflicting reports as to the reputation of drugs among experts in the field is not a matter well left to a court without chemical or medical background.' The determination whether a drug is generally recognized as safe and effective within the meaning of § 201(p)(1) necessarily implicates complex chemical and pharmacological considerations. Threshold questions within the peculiar expertise of an administrative agency are appropriately routed to the agency, while the court stays its hand."

Applying these principles here, we hold that the district court erred in assuming the task of deciding whether Insulase is safe and effective rather than limiting itself to determining whether on the basis of substantial evidence Insulase is generally recognized to be safe and effective. Had the district court applied the proper standard to the undisputed record facts, it could only have reached the conclusion — which the court had reached in denying preliminary injunctive relief — that Insulase 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 labelling thereof.

There is no published scientific literature with respect to Insulase enabling qualified experts to make the necessary determination. Indeed, the scientists called as expert witnesses at trial had learned of Insulase only in connection with their preparation for trial. The studies offered as the basis for their opinion and essential to the determination of safety and effectiveness were made by Premo and not available to the community of experts generally. Even on the basis of these limited studies, there was no general consensus among the experts but a sharp difference of opinion between them, both as to the methods used and the results claimed. Lastly, although Premo has produced and sold at wholesale some 16,500,000 Insulase tablets (some of which have been seized in Government actions under 21 U.S.C. § 334), there is no evidence that Insulase has been used to a material extent or for any substantial period of time.

The reasons given by the district court for not following the "general recognition" standard prescribed by 21 U.S.C. § 321(p) is that it "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," 475 F.Supp. at 55. We disagree. The conditions fixed by Congress in 21 U.S.C. § 321(p) permitted many drugs generally recognized as safe prior to the enactment of the Act in 1938 to be marketed without filing NDAs. Similarly, when the Act was amended in 1962, many drugs then on the market and generally recognized as safe and effective were permitted to be marketed under § 321(p) without the filing of an NDA. And even those drug products were subject to review by panels of the National Academy of Sciences/National Research Council for evidence of the drug product's efficacy. On the other hand, later developed "me-too" products such as Insulase are required to apply for FDA approval for the undisputed reason that a difference in inactive ingredients, as exists here, when combined with the active ingredient, can affect the safety and effectiveness of the drug product. Although the FDA has sought to minimize the cost and delay in obtaining approval for "me-too" drug products by permitting the proposed marketer to file an ANDA instead of an NDA, the purpose of the Act is to subject all such drug products not generally recognized as safe and effective (whether or not labelled "me-too" products) to the premarket clearance requirements of the Act.

Since Insulase does not meet the "general recognition" standard prescribed by 21 U.S.C. § 321(p) and is therefore a "new drug" within the meaning of the Act, the judgment of the district court must be reversed and the case remanded with directions to dismiss the complaint.