Six and a half years after the Food and Drug Administration began its bumbling efforts to withdraw approval of the New Drug Application (NDA) for Alevaire, an aerosol prescription drug administered to patients with chronic respiratory diseases, on the ground of lack of substantial evidence that it will have the effect it is represented to have under the conditions of use set out in its labeling, the case is back again--only a few months after we thought we had sent it on its way to agency disposition, Sterling Drug, Inc. v. Weinberger, 503 F.2d 675 (2 Cir. 1974). On this occasion it is here on appeal from a decision by Judge Pierce in the District Court for the Southern District of New York denying an injunction and dismissing the complaint in an action by the plaintiff manufacturers and distributors for a declaratory judgment declaring that the FDA's proposal of August 1, 1974, to withdraw approval of the NDA pursuant to a notice of opportunity for hearing published on August 13, 1974, 39 F.R. 29013, was barred by res judicata in part and illegal in remaining part and for an injunction to prevent the FDA from withdrawing approval of the NDA. We affirm, although, with respect to the res judicata argument, on an analysis differing from that of the district court.

We shall not repeat the ancient and medieval history of FDA and court proceedings relating to Alevaire which are recounted in Judge Bryan's opinion of last May, 503 F.2d 675, but will confine the narration to subsequent developments. The notice of August 13, 1974, after summarizing the prior proceedings, went into a discussion substantially as follows: 'It is possible to conclude that Alevaire is a combination product, with several ingredients intended to contribute to the overall claimed therapeutic effect.' The chief basis for this is that 'in its labeling the sponsor has at times indicated adherence to this interpretation.'1 'On the other hand, it is also possible to consider Alevaire as a single active component, tyloxapol, with a vehicle that happens to have some activity of its own in the treatment of patients with difficulty in mobilizing pulmonary secretions,' as indicated by recent packaging.2 It is by no means surprising that the same product could be either a 'fixed combination' or a 'single entity' drug, depending upon how it is characterized by its manufacturers and distributors, since the statutory authority underlying these FDA actions lays stress on representations of effectiveness made in the drug's labeling.3 How the product is categorized has important bearing, however, on the nature of the 'adequate and well-controlled investigations,' 21 U.S.C. § 355(d), necessary to sustain the burden of 'substantial evidence that the drug will have the effect it purports or is represented to have,' 21 U.S.C. § 355(e), and thus justify a hearing. If Alevaire is a combination, demonstration of effectiveness 'would require, at a minimum, a study including groups treated with water alone, water plus bicarbonate, water plus tyloxapol, water plus bicarbonate plus tyloxapol.'4 See 21 C.F.R. 3.86 & 314.111(a)(5)(ii)(a)(4). On the other hand, if Alevaire is a single active component in a several component vehicle, the evidence need only show 'that Alevaire is more effective than its admittedly active vehicle.' See 21 C.F.R. 314.111(a)(5)(ii)(a)(4). This would demand only 'a two-group trial: patients treated with the vehicle (water, bicarbonate, and glycerin) alone vs. patients treated with Alevaire.'5 The Bureau of Drugs was presently of the view, based in part on the pre-1970 labeling, that Alevaire is a fixed combination drug product. Any request for a hearing should either provide evidence of effectiveness as a fixed combination drug product or demonstrate that Alevaire is a single entity drug product6 and is effective as such. Failing the submission of evidence warranting a hearing, the Bureau proposed to withdraw approval of the NDA on the ground of ineffectiveness. Requests for a hearing had to be filed by September 12 and supporting evidence by October 15, 1974. See 21 C.F.R. 314.200. Plaintiffs requested such a hearing on September 10 but did not follow up by submitting data to support that request.

Instead the manufacturers began this action in the district court on September 30, 1974, against the Secretary of Health, Education and Welfare and the Commissioner of Food and Drugs. The theory of the complaint was as follows: By the proceedings culminating in our last decision, the defendants had formally conceded and abandoned any claim that the effectiveness of Alevaire as a single component drug can only be tested against its vehicle, and relitigation of that issue was barred on grounds of res judicata and/or collateral estoppel. With respect to the claim that Alevaire was a fixed combination drug and ineffective as such, the notice published on August 13 showed that the FDA would be unable to meet the 'new information' test prescribed by § 505(e)(3) of the Food, Drug and Cosmetics Act, 21 U.S.C. § 355(e)(3).7 Plaintiffs sought a declaration that defendants were barred by res judicata and/or collateral estoppel from withdrawing approval of the NDA for Alevaire on the ground that effectiveness must be demonstrated by comparing Alevaire to its vehicle and that they were barred by statute from proceeding against Alevaire as a fixed combination. Plaintiffs also moved for preliminary injunctive relief to prevent defendants from proceeding with the August 1 proposal to withdraw approval of the NDA and for permanent injunctive relief to prevent defendant's action on that or any other like proposal.

On October 1, Judge Pierce ordered defendants to show cause why they should not be temporarily enjoined from proceeding on their proposal to withdraw approval of the NDA during the pendency of this action. The order was returnable on October 10, on which day the order was consolidated with a trial on the merits and that trial completed. On October 23, Judge Pierce signed a stipulation in which the parties agreed that plaintiffs would be given an extension of time during which to file data in support of their request for a hearing from October 15 until 15 days after the district court's decision on the motion for a preliminary injunction.

Judge Pierce rendered an opinion on October 31 denying the motion for an injunction and dismissing the complaint. He concluded that he need not reach the merits of plaintiffs' arguments since the complaint must be dismissed for failure to exhaust administrative remedies. On November 12, 1974, a panel of this court granted an injunction pending the hearing of an expedited appeal; at argument on December 18 this panel extended the injunction pending decision.

Plaintiffs' basic position is that we have entered a brave new world with respect to injunctions against the conduct of administrative proceedings. We are told in effect that the warning against judicial intrusion in Myers v. Bethlehem Shipbuilding Corp., 303 U.S. 41, 58 S. Ct. 459, 82 L. Ed. 638 (1938), long regarded as the leading case in this area, is old-hat and that Leedom v. Kyne, 358 U.S. 184, 79 S. Ct. 180, 3 L. Ed. 2d 210 (1958), has become the seminal decision. Plaintiffs seemingly are not concerned with the Supreme Court's statement that:

The Kyne exception is a narrow one, not to be extended to permit plenary district court review of Board orders in certification proceedings whenever it can be said that an erroneous assessment of the particular facts before the Board has led it to a conclusion which does not comport with the law.

Boire v. Greyhound Corp., 376 U.S. 473, 481, 84 S. Ct. 894, 899, 11 L. Ed. 2d 849 (1964). They are likewise not discouraged over the fact that the two cases cited by them for the proposition that 'Leedom has been increasingly followed and used to check illegal agency action,' namely Boire v. Miami Herald Publishing Co., 343 F.2d 17 (5 Cir.), cert. denied, 382 U.S. 824, 86 S. Ct. 56, 15 L. Ed. 2d 70 (1965), and FTC v. J. Weingarten, Inc., 336 F.2d 687 (5 Cir. 1964), cert. denied, 380 U.S. 908, 85 S. Ct. 890, 13 L. Ed. 2d 796 (1965), reversed injunctions granted by district courts. Nothing advanced by plaintiffs persuades us to depart from the general position we took after full analysis, only two years ago, that the most that can be extrapolated from Leedom v. Kyne is that an injunction may be issued 'if an agency refuses to dismiss a proceeding that is plainly beyond its jurisdiction as a matter of law or is being conducted in a manner that cannot result in a valid order,' Pepsico, Inc. v. FTC, 472 F.2d 179, 187 (2 Cir. 1972) (footnote omitted), cert. denied, 414 U.S. 876, 94 S. Ct. 44, 38 L. Ed. 2d 122 (1973)--a 'possibly over-generous' standard we found not to have been met by the plaintiff there.8 Until otherwise instructed we shall continue to regard Myers as stating the general rule and Leedom v. Kyne as being what the Supreme Court says it is--a narrow exception.

Relying on the per curiam reversal of Otis & Co. v. SEC, 85 U.S.App.D.C. 122, 176 F.2d 34, in SEC v. Otis & Co., 338 U.S. 843, 70 S. Ct. 89, 94 L. Ed. 516 (1949), defendants would apparently contend that even our Pepsico formulation goes too far in permitting an injunction when a plaintiff relies on a claim of res judicata and that such a claim can never constitute a sufficient basis for enjoining a new administrative proceeding. It is hard to see why this should be so as a matter of principle, and we would be reluctant to extract so broad a proposition from a per curiam reversal made on a petition for certiorari without oral argument and supported by three citations, none of which dealt with res judicata.9 The Otis case involved the effect of a prior judicial determination, SEC v. Harrison, 80 F. Supp. 226 (D.D.C.1948), motion to vacate denied, D.C.Cir., Feb. 21, 1950 (unreported), rehearing denied, 87 U.S.App.D.C. 232, 184 F.2d 691 (1950), vacated as moot, 340 U.S. 908, 71 S. Ct. 290, 95 L. Ed. 656 (1951), not on the ultimate issue before the SEC--expelling Otis from the National Association of Securities Dealers and revoking its broker-dealer registration--but on whether there was sufficient evidence of fraud to justify piercing the attorney-client privilege and thus securing testimony in an investigative proceeding from attorneys who allegedly had acted at Otis' bidding to bring a suit that freed Otis from what had become a burdensome underwriting. The Court's decision could have rested on the basis that, even without the attorneys' testimony, in such an administrative proceeding10 the SEC might hold the circumstances to have been sufficiently suspicious to transfer the burden of going forward to Otis.11 The very court of appeals that had been reversed in the Otis case has not given it so expansive a reading. Elmo Division of Drive-X Co., Inc. v. Dixon, 121 U.S.App.D.C. 113, 348 F.2d 342 (1965).12

We thus prefer to rest our ruling with respect to the 'single active component' theory, the part of the FDA's case claimed to be barred by res judicata, not on the broad interdiction claimed by defendants but on a different ground relating to the facts of this case. This is that '(o)nly a final decision can be res judicata.' 2 Davis, Administrative Law Treatise § 18.06, at 584 (1958). We have searched the record in vain for a determination by the FDA or a reviewing court that Alevaire had been proved to be effective as a single active component drug. The reinstatement of the NDA on June 14, 1973, 38 F.R. 15861, see 503 F.2d at 679, clearly was not such a determination; it stated merely that, upon reviewing the petition for reconsideration, the FDA had concluded that the requests for a hearing should be reevaluated under 21 U.S.C. § 355(f) & (h). Compare Jenkins v. Macy, 357 F.2d 62, 66--67 (8 Cir. 1966). Neither can we find such a determination in the August 7, 1973, notice of withdrawal of approval, 38 F.R. 21515, see 503 F.2d at 679--680. Apart from the point that this notice has been set aside at plaintiffs' request, a determination that Alevaire had not been shown to be effective as a fixed combination drug, which the FDA considered it to be on the basis of the labeling used when the NDA became effective and appearing in the Physicians' Desk Reference book as late as 1971, did not preclude a fallback to the position that Alevaire had also not been proved effective as a single active component drug if that was what, as plaintiffs now seem to contend, it really is. When this court set aside the August 1973 notice of withdrawal on plaintiffs' claim that they had not been given adequate notice of the specific grounds on which the FDA proposed to withdraw approval (lack of substantial evidence that the product was effective as a fixed combination drug) and indicated doubts about the FDA's ability to prevail on that basis, 503 F.2d at 683, the FDA was entitled to proceed in the alternative and advance again the theory that Alevaire was not effective as a single active component drug, since that issue had never been adjudicated. Compare FTC v. Raladam Co., 316 U.S. 149, 62 S. Ct. 966, 86 L. Ed. 1336 (1942); NLRB v. Zimnox Coal Co., 336 F.2d 516 (6 Cir. 1964). The case would stand differently if, during the period when the FDA was proceeding on the single active component theory, it had found that the studies submitted by plaintiffs had established Alevaire to be effective as such; in that event the FDA would be bound to show some new evidence that would call its earlier decision into question before a new proceeding on that theory could be justified. Compare H.S.D. Co. v. Kavanagh, 191 F.2d 831, 846 (6 Cir. 1951), and 2 David, supra, § 18.03, at 562. But the most that the FDA had ever determined on that score was that plaintiffs' petition for reconsideration and the materials supporting it justified reevaluation of their request for a hearing on the efficacy of Alevaire as a single component drug, 38 F.R. 15861, and later, in the order which we vacated, that there was no need to decide that question since it its view Alevaire was a fixed combination drug, 38 F.R. 21515.13 This would hardly create res judicata in private civil litigation, much less under the 'qualified or relaxed set of rules', see 2 Davis, supra, § 18.03, at 568, properly applicable to proceedings by an agency charged with responsibility to protect the public against ineffective drugs.

There is even less merit in plaintiffs' request that the FDA should also be enjoined from proceeding on the theory that Alevaire is ineffective as a fixed combination drug--exactly what this court said it might do, 503 F.2d at 683. The argument is that the August 1974 withdrawal notice is predicated only on old information, to wit, the former labeling and the NAS-NRC study, and not on the 'new information . . . evaluated together with the evidence available . . . when the application was approved', which the statute requires, see note 3 supra. Although we need not decide the point, there appears to be merit in defendants' contention that 'newness' dates from the time of the grant of the NDA rather than from the date of the last previous agency action. At least that is an entirely reasonable reading of 21 U.S.C. § 355(e)(3) and the FDA can scarcely be so hobbled by its statutory authority that after initial use of an NAS-NRC study it is thereafter barred from relying upon it again. Compare Bell v. Goddard, 366 F.2d 177 (7 Cir. 1966). At minimum we agree with Judge Pierce that the FDA's contentions on this branch of the case are sufficiently meritorious that they should be initially determined by the agency. There is no such apparent, indeed, conceded, action in violation of the statute as in Leedom v. Kyne, supra, 358 U.S. at 187, 79 S. Ct. 180, 3 L. Ed. 2d 210.

Plaintiffs have requested that, in the event of affirmance, we allow them a still further period for furnishing the data, information and analysis relied on to justify a hearing, which the August 1, 1974, order had required to be filed by October 15, 1974. When the motion for a stay was before us, we warned plaintiffs' counsel that he should not assume that further extensions would be granted but should begin the preparation of the necessary studies. While we would thus be abundantly justified in granting only the shortest of further extensions, we think it better to allow some additional time rather than create a new ground for procedural argument. We shall therefore extend the date for filing the information until February 19, 1975. This date will stand, save as it may be extended by the FDA, regardless of any further petitions that plaintiffs may file--unless, of course, one of these should be granted.

Affirmed.