Unlike the drug manufacturers in Ben-tex,1 this plaintiff markets a line of citrus bioflavonoid drugs,2 of which all but two were covered by NDAs issued at various times in 1955 and 1956. Like the plaintiffs in Bentex, however, it seeks by an action for declaratory judgment to secure the benefit of the exemption available under the “grandfather clause”3 from the enlarged definition of a “new drug” included in the 1962 Amendments to the Federal Food, Drug, and Cosmetic Act of 1938. The defendants, who are the Secretary of Health, Education and Welfare (hereinafter referred to as HEW) and the Commissioner of the Food and Drug Administration (hereafter referred to as Commissioner), urge that jurisdiction should be refused on two grounds: 1. Primary jurisdiction lies with HEW; and 2. Failure to ex*226haust administrative remedies. They, also, attack the right of the plaintiff to claim the exemption. The District Court sustained jurisdiction and, largely on the basis of a Stipulation of Facts, upheld plaintiff’s right to the statutory exemption both for its NDA’d and its non-NDA’d drugs. We reverse.

The threshold question raised by the defendants and overruled by the District Court may be quickly disposed of. Under similar circumstances in Bentex, we sustained the right of the District Court to entertain an action for declaratory judgment. We reach the same result here. Since we dismiss the claim of the plaintiff for exemption on behalf of its drugs on substantive grounds, it is unnecessary to consider the additional objection that plaintiff has failed to exhaust administrative remedies.

The substantive issue posed by this action is the right of the plaintiff to the exemption provided by section 107 (c) (4) from the revised definition of “new drug” incorporated in the 1962 Amendments. In resolving that issue, we must differentiate, even as the “grandfather clause” itself does, between the plaintiff’s drugs, which were covered by an “effective NDA”,4 and those, which were marketed without an NDA. The Act makes a distinction in “grandfather” rights between a drug marketed under an NDA5 and one marketed without an N DA. In the case of a drug covered by a previously approved NDA, the 1962 Amendments required the Secretary to withdraw the approved NDAs if after notice and opportunity of hearing, the applicant failed to file substantial evidence,6 that the drug previously approved is both safe and effective.7 For such drugs, however, a grace period or temporary “grandfather right” was granted. Under it, the manufacturer was given two years within which to develop his showing of effectiveness and, during this period, the Secretary was prohibited from withdrawing or suspending the previously granted NDA.8 On the other hand, a non-NDA’d drug which met the criteria stated in section 107(c) (4) was exempted permanently from the amended definition of “new drug” made by the 1962 Amendments and was thereby relieved of securing an approved NDA as a condition for marketing clearance. The statutory criteria for this permanent “grandfather” exemption are stated as “any drug which, on the day immediately preceding the enactment date, (A) was commercially used or sold in the United States, (B) was not a new drug as defined in section 201 (p) of the basic Act as then in force, and (C) was not covered by an effective application under section 505 of that Act”.

It is the contention of the plaintiff that all its drugs in question, both those previously NDA’d and those not, are protected by the permanent “grandfather clause” (i. e., Section 107(c) (4)). Because the statute seemingly makes a distinction between the two, it is proper to consider separately the two groups of drugs: i. e., those having NDAs and those without NDAs.

Taking up first plaintiff’s NDA’d drugs: There is no dispute that such drugs met criteria (A) and (B), as set forth in the “grandfather clause”, but the defendants seriously dispute the claim that they meet condition (C). Facially at least, this contention of the defendants seems unanswerable. These drugs are “covered by an effective application” or NDA, and are thus specifically barred by condition (C) from qualifying for exemption from the ap*227plication of the effectiveness Amendments of 1962. The District Court found, however, that before “the day immediately preceding the enactment date”, which was October 9, 1962, the previously granted NDAs had been effectively and practically withdrawn and that accordingly the drugs were not covered by an effective NDA on the crucial date of October 9, 1962. The error in this reasoning, however, is that it assumes that a manufacturer may effect a withdrawal of an effective NDA, either by a formal notice or by discontinuing compliance with the reporting requirements for NDA’d drugs. While an applicant may, during the pendency of his application, withdraw his application, he has no such right after approval of the application by the Secretary. At that point only the Secretary can withdraw the approval. As one commentator has accurately summed up, “It is true that a manufacturer may withdraw a pending NDA. Sec. 21 C.F.R. sec. 130.8 (1971). However, no provision in the law permits a manufacturer to withdraw an effective NDA; only the FDA can do so through Section 505(e) procedure”.9 Prior to October 9, 1962, there was in this case no proceeding by FDA under Section 505(e) with reference to plaintiff’s NDA’d drugs and there was accordingly no valid withdrawal of the plaintiff’s effective NDAs, on or before the enactment date of the 1962 Amendments.

The plaintiff, though, presses another theory upon the basis of which it claims the previously issued NDAs are to be regarded as ineffective on October 9, 1962. Thus it argues that its pre1962 NDA’d drugs became generally recognized as safe on or before October 9, 1962. So much the defendants seem to concede in the Stipulation of Facts submitted to the District Court. From this fact, it reasons that its NDA drugs ceased to be “new drugs” as defined in the Act, on or before October 9, 1962, and, ergo, its previously issued NDAs were no longer needed or “effective” on the critical date of October 9.10 The difficulty with this argument, plausible though it may be, is that it would make surplusage of requirement (C) in the exemption statute. Thus, if a drug met the test set up in (B), that is, was generally recognized as safe on October 9, 1962, it would not be necessary, under the plaintiff’s argument, to consider whether (C) was applicable or not. Such a construction of the exemption statute, under which a clearly stated condition to its application is to be treated as a nullity, offends the well-settled rule of statutory construction that all parts of a statute are to be given effect if at all possible.11 It is manifestly possible to give effect to the conditions enunciated in both (B) and (C). There are many drugs that satisfy both conditions, that is, are generally recognized as safe and effective and are being marketed without an approved NDA. There is nothing inconsistent in the two requirements. Moreover, condition (C) represented a limitation on the right to an exemption that the Congress clearly and unmistakably intended to apply. The Congress never intended that a drug being marketed under an approved NDA might qualify under the “grandfather clause.” This is plain from the comment in the Conference Committee Report that the exemption was to apply “to existing labeling claims o.f drugs that have never previously been subject to the new-drug procedure”. (Italics added.) Conference Rep. 2526, 87th Cong., 2d Sess., p. 23. Moreover, the argument of the plaintiff would run counter to the principle that statutory exemptions, particularly as applied to statutes concerned with public *228health and safety, are to be strictly and narrowly construed.12

The plaintiff has, however, two drugs,13 involved in this proceeding, which were generally recognized as safe and were14 marketed as “old drugs” without an approved NDA on October 9, 1962. These drugs, as we earlier indicated, present separate problems from those drugs for which there are approved NDAs. They fall into the category of what are generally described in the trade as “me-too” drugs.15 Such a drug, if generally regarded as safe on October 9, 1962, meets literally the criteria for exemption stated in the "grandfather clause”. To sustain the exemption, however, creates an inequitable result, provided the pioneer drug was NDA’d. In that event, the pioneer drug would be subject to withdrawal of marketing privilege absent substantial evidence of effectiveness, whereas its copy would enjoy immunity from any such requirement under Section 107(c) (4). Most commentators, while admitting the incongruity of this result, justify it as one compelled by the literal language of the statute.16 Their reasoning is similar to that of the Court in Pfizer, Inc. v. Richardson (2d Cir. 1970) 434 F.2d 536, 542, where speaking to a somewhat illogical provision in this same Act, Judge Friendly said: “A sufficient answer is the simple if not altogether satisfying one that Congress said so.”17 The FDA itself has recognized the vexing problem presented by the “me-too” drugs and has sought to resolve it by a change in its position on the scope and application of an NDA.

It is the contention of the FDA that an approved NDA covers not merely the marketing of the parent but also its “me-too” offsprings and for that reason the “me-too” drugs have been permitted to be marketed without an NDA. Accordingly, under this theory, the withdrawal of the approved NDA of the pioneer operates as a withdrawal of marketing rights for the “me-too”, unless the latter, either individually or in conjunction with its pioneer, provides substantial evidence o,f effectiveness. This view has, however, been severely criticized and with considerably reason. It is, as one critic has observed, “at variance with the uniform position it (FDA) has taken over the years with regard to the nature of NDAs.” This position, which is termed the “personal approach” holds that “Section 505 applies to drugs as individual articles, not as collective groups, and that each manufacturer of a new drug must file his own NDA.” This critic concludes with the observation that it is “an unjustifiable exercise in semantics to say that a drug legally marketed without an NDA was *229‘covered’ by the NDA of another manufacturer’s drug.” 18

That the policy of FDA has heretofore been contrary to the position now taken by it is further illustrated by the circumstances under which at least one of the “me-toos” of the plaintiff began marketing. Prior to marketing Bivam, one of its “me-toos” similar in formula to other drugs previously NDA’d by it, the plaintiff inquired of FDA whether it was an “old drug” entitled to be marketed without an NDA. FDA, after reviewing its composition and labeling, advised the plaintiff it was a product “generally regarded as safe” (and thus an “old drug”) and could be marketed without an NDA. There was no suggestion by the plaintiff that it sought to market this drug under any previous N DA granted one of its products nor did the FDA base its advice on that basis. Both the plaintiff and FDA assumed at that time that a “me-too”’ drug, which had become generally recognized as safe, was entitled to be marketed without an NDA; in short, that the qualification for marketing a “me-too” drug was general recognition of safety and not the NDA of its pioneer.

It would seem that the consistent construction of the Act by the FDA for thirty years19 and a construction which accords with the literal language of the Act itself may only be changed by Congress itself.20 In fact, the General Counsel of FDA, in testimony before a House Subcommittee Hearing on Drug Efficacy, Part 2 (91st Cong., 1st Sess., 1969) p. 375, expressed the wish that Congress would “pass” a clarifying amendment on this issue, conceding that the position of his agency was in considerable doubt.

But even if it be assumed that “me-too” drugs are generally entitled to section 107(c) (4) protection, provided they were generally recognized as safe on October 9, 1962, that does not resolve the right of the plaintiff’s “me-toos” to exemption. As has been pointed out, the reasoning on which “me-toos” are regarded as not covered by the NDAs granted the manufacturers of their pioneers is that an NDA is regarded as “personal” to the manufacturer submitting the application and to the drug covered. But in this case, the “me-toos” are similar in formula and labeling to other drugs for which the plaintiff itself applied and obtained NDAs. It is true that, in the case of one drug at least, to which reference has already been made, plaintiff’s “me-toos” were regarded as exempt, not because plaintiff had an approved NDA for a similar drug,, but because FDA was of the opinion that it met the requirements for classification as an old drug. Nonetheless, it is the “personal” character of the NDA that has been deemed as the basis on which it is contended that the “me-toos” are not covered by the NDA granted another manufacturer, albeit the drugs involved may be similar. That reasoning manifestly cannot sustain a right of exemption in favor of plaintiff’s “me-toos”. The plaintiff’s NDAs, being “personal” to it, would cover all its products similar in formula, including those specifically described in its applications and all others like in *230formula. The similarity in formula, between plaintiff’s NDA’d drugs and its “me-toos” is stipulated. Under those circumstances, both the NDA’d and the “me-too” drugs will be treated alike and neither can qualify for exemption under the terms of section 107(c) (4). It is recognized that this conclusion places the plaintiff in a less favorable position than that occupied by others who may have copied its product prior to October 9, 1962. That inequity is, however, inherent in the law and may only be redressed by Congress, not by the Courts under the guise of construction.

Reversed, with directions to enter judgment for the defendants.

Reversed.