The United States appeals from Judge Mishler's order granting in part a motion by Diapulse Corporation of America ("DCA") to modify a permanent injunction in order to allow it to market a modified version of its Diapulse machine. DCA cross-appeals from that part of Judge Mishler's order that continues to prohibit DCA from marketing the original Diapulse device without prior Food and Drug Administration ("FDA") approval. We affirm.

This case arises out of one of many attempts by DCA to market a device for which some demand exists in the medical community but which operates on scientific principles rejected by the FDA. Judge Mishler concluded that DCA must be allowed to market a modified version of the device because in 1973 the FDA granted another company approval to market a similar device.

A detailed review of the facts is unfortunately necessary. Medical science has long recognized the usefulness of heat in treating a variety of ailments. Superficial heat can be applied through hot towels, heating pads, and similar devices, while deep heat can be applied through diathermy machines. Diathermy machines create an electromagnetic field, which induces alternating currents and heat inside a patient's tissues. A diathermy machine, when in use, produces a continuous field, and therefore the average power it delivers is one-half of its peak power.1 The amount of heat it delivers to the patient is by definition equal to the average power.

Some doctors have concluded that the flow of electric current in the body produces certain beneficial effects entirely apart from the heat it generates. The precise mechanisms by which these effects are produced are unimportant for present purposes, but DCA and some doctors claim that electric fields in human tissue stimulate certain realignments within cells that can promote healing.2 We will refer to these realignments as athermal electric effects in order to distinguish them from the healing effects produced by heat.3

DCA markets in many other countries a machine called the Diapulse, which produces athermal electric effects. Physically, the Diapulse resembles a diathermy machine and also creates an electromagnetic field that can induce current in human tissue. However, it produces stronger currents but less heat than a diathermy machine by adding a stage called a pulse modulator. The pulse modulator creates a field that rises to its maximum intensity in short bursts, or pulses, the field being weak to nonexistent most of the time, in contrast to the continuous field of a diathermy machine.4 During each pulse the field strength and, therefore, the delivered power of the Diapulse machine, is much greater than that of a diathermy machine (975 watts peak power for Diapulse on its highest setting versus 200 watts peak power for a typical diathermy).5 However, the Diapulse bursts of current are very brief, occurring from 80 to 600 times per second depending on the setting and lasting for 65 millionths of a second each. Therefore, it delivers power for only a small fraction of each second and the maximum average power--or heat--delivered by a Diapulse machine is less than that of a typical diathermy machine (38 watts versus 100 watts). Were the Diapulse machine continuous-wave--i.e., delivering power constantly rather than in short bursts--the strength of the current induced would create so much heat as to seriously damage human tissue.

The virtue, if any, of Diapulse's pulsed field is that it maximizes the supposed athermal electric effects while minimizing heat. Proponents of the Diapulse also claim that the frequency of the pulse bears a close relation to the frequency at which nerve endings discharge, and the Diapulse machine therefore produces a beneficial stimulation of nerve fibers. Neither of these effects has anything to do with heat, and neither would be useful to anyone who wanted only heat treatments. Indeed, the Diapulse machine produces negligible heat.

Coming finally to the crux of the matter, the FDA has rejected the claims of therapeutic value made on behalf of athermal electric effects. Accordingly, on December 17, 1965, the FDA filed a seizure action against a Diapulse device pursuant to 21 U.S.C. Sec. 334. The FDA claimed the machine was mislabelled in violation of 21 U.S.C. Sec. 352(a) because the labelling falsely represented that the machine was effective in treating a variety of conditions. A jury agreed with that claim, and judgment was entered for the government on March 31, 1967. We affirmed. United States v. Diapulse Manufacturing Corp., 389 F.2d 612 (2d Cir.), cert. denied, 392 U.S. 907, 88 S. Ct. 2059, 20 L. Ed. 2d 1365 (1968).

DCA nevertheless continued to market the Diapulse, and in April, 1968, the FDA sought and was eventually granted a preliminary injunction. We again affirmed, United States v. Diapulse Corp., 457 F.2d 25 (2d Cir.1972), and Judge Rosling entered a permanent injunction on July 18, 1972.

The injunction of course did not end the demand for a Diapulse-like machine among doctors who disagreed with the FDA's evaluation of athermal electric effects. This demand induced a former Diapulse employee, doing business as United Medical Equipment Co. ("UME"), to develop a device called a Magnatherm. Because it was roughly identical to the Diapulse, the FDA responded with a seizure action against the Magnatherm.

UME, however, hit upon a strategy that pacified the watchdogs at the FDA. It modified the Magnatherm so as to increase the amplitude (height) and width of the pulses and to permit the user to vary the pulse frequency. At the higher pulse frequencies, the machine would deliver power for a sufficiently long fraction of each second to generate some non-negligible heat in human tissues. However, it continued to produce a pulsed output and retained the lower pulse frequency settings similar to the original Diapulse device. The Magnatherm, then, had essentially two capabilities. First, it could raise deep tissue temperature to 104?F in twenty minutes or less at some frequency settings. This satisfied FDA's then-existing standards for medical diathermy. Second, it could produce athermal electric effects. Notwithstanding the latter capability, however, UME agreed to market the modified Magnatherm solely as a heat producing diathermy machine, and to make no claims regarding athermal electric effects. In January, 1973 the FDA permitted UME to market the Magnatherm on the basis of that agreement.

It must be emphasized that the Magnatherm is not mechanically superior to a conventional diathermy machine as a heat producing healing device. Moreover, it contains complex and presumably costly features superfluous to its heat producing function. FDA approval would seem, therefore, to be commercially irrelevant since purchasers interested only in the heat producing function would always choose a conventional diathermy machine. Indeed, at oral argument counsel for the government stated flatly that only members of the athermal electric effects "cult" would choose a Magnatherm over a conventional diathermy machine.

However, UME appears to have valued FDA approval because it was confident that doctors who believed in the therapeutic value of athermal electric effects would recognize, without the benefit of advertising, the modified Magnatherm's full capabilities. Meanwhile, DCA understandably attempted to market what it called a "P/emf modification kit," capable of modifying a Diapulse machine in roughly the same manner as UME had modified its Magnatherm machine. The FDA, despite its approval of the Magnatherm, instituted a criminal contempt action against DCA for marketing the P/emf kits.

The district court, now Judge Dooling, dismissed the contempt action on the ground that the government had not proved criminal intent. However, at FDA's request, Judge Dooling modified the permanent injunction to prohibit DCA from marketing any high frequency electromagnetic device unless: (i) the peak power available at any setting of the device does not exceed twice the average power at that setting;6 (ii) the device can raise the temperature of human tissue at a depth of two inches in the thigh muscle to 104?F in twenty minutes or less at a majority of its operational settings; and (iii) the device can raise the temperature of such tissue to a range of temperatures from 104?F to 113?F at a series of settings. This injunction is still in effect and is the subject of the present proceeding.

Condition (i) of the injunction essentially precludes any redesign of the Diapulse that would use a pulsed electromagnetic field because a continuous alternating field can deliver at best an average power that is exactly one-half the peak power, see note 1, supra. Because a pulsed field by definition delivers power for only a small fraction of each second, it cannot maintain an average power that is one-half the peak power.7 The net effect of the modified injunction is thus to require that any redesigned Diapulse be virtually identical to an ordinary diathermy machine and that, therefore, DCA abandon the market of believers in athermal electric healing effects.

Because the Magnatherm machine also produces a pulsed field, it cannot meet condition (i). In the FDA's view, it also does not meet condition (iii). However, the injunction runs only against DCA, and UME continues to market its device.

After Judge Dooling modified the injunction, DCA continued to seek to convince the FDA and the courts of the efficacy of athermal electric effects, see, e.g., United States v. An Article of Device ... Diapulse, 650 F.2d 908 (7th Cir.1981), and to obtain FDA permission to market a machine called the P/emf with labelling similar to that of the Magnatherm. The FDA denied these requests, and DCA moved the district court, now Judge Mishler, to modify the 1974 injunction to allow it to market the original Diapulse device in light of allegedly new scientific evidence, or to permit it to market the P/emf. On October 27, 1981, the district court denied the motions and remanded to the FDA to determine the legality of the P/emf labelling, in light of the P/emf's similarity to the Magnatherm.

The FDA rejected DCA's scientific evidence relating to the beneficial qualities of athermal electric effects and again disapproved marketing of the Diapulse. It also denied approval for DCA's proposed labelling for the P/emf since DCA's data did not establish that the P/emf met the heating standards required by the 1974 injunction, which are equivalent to FDA's current standards for medical diathermy.8 With regard to the question of the P/emf's similarity to the Magnatherm, the FDA stated: (1) it had approved the Magnatherm under its pre-1974 diathermy standards, which were no longer applicable; and (2) the pulse width and frequencies of the P/emf differed from those of the Magnatherm.

DCA returned to the district court seeking modification of the 1974 injunction. Judge Mishler held that FDA had acted arbitrarily and capriciously in approving the Magnatherm while rejecting the relabelled P/emf, because it had not explained how the differences between the two machines affected their relative effectiveness as heat producing devices. He also rejected FDA's contention that the differences between its pre-1974 and post-1974 diathermy standards justified disparate treatment of the Magnatherm and P/emf. With regard to FDA's contention that the late Judge Dooling fashioned the modified injunction with full knowledge of its disparate impact on the Magnatherm and P/emf, Judge Mishler concluded that the record did not reveal that Judge Dooling's knowledge of the Magnatherm's capabilities and of FDA's permissiveness toward it was sufficient to alert him to the problem. Finally, Judge Mishler found that FDA had displayed "a pattern of bias" against DCA. He then modified the 1974 injunction to permit DCA to market the P/emf, under a new name, with labelling identical to that used on the Magnatherm. Nevertheless, he refused to allow DCA to market the original Diapulse. This appeal and cross-appeal followed. We affirm.

We agree with Judge Mishler that the record does not demonstrate that Judge Dooling knew of the similarity of the Magnatherm and P/emf machines when he modified the injunction against DCA in 1974. We also regard the fact that the FDA has permitted UME to market a machine which does not meet the requirements of the 1974 injunction obtained by the FDA against DCA as a sufficient cause to warrant judicial reexamination of that decree. See System Federation v. Wright, 364 U.S. 642, 647, 81 S. Ct. 368, 371, 5 L. Ed. 2d 349 (1961).

In response to Judge Mishler's remand for consideration of the inconsistencies in its treatment of the two machines, the FDA noted the revision of diathermy standards and the differences in pulse widths of frequencies. Neither claim is sufficient to justify continuation of the 1974 injunction.

First, neither claim is at all responsive to the provision of the injunction that effectively bars the marketing of any machine using a pulsed field. In seeking that provision, the FDA sought to bar DCA from marketing any device which might be used to produce athermal electric effects, notwithstanding its permissive attitude toward the Magnatherm. That portion of the decree simply cannot be reconciled with the FDA's treatment of the Magnatherm machine.

The response is also inadequate in other ways. It is true that the operating characteristics of the two machines are different but not in a way relevant to the present issue. The modified Magnatherm's pulses are wider and, at each machine's highest frequency settings, closer together than those of the P/emf. However, these differences merely enable the Magnatherm to deliver power for a slightly greater portion of each second than the P/emf and thus to have a slightly better average-power-to-peak-power ratio than the P/emf. At best, this means only that the Magnatherm is slightly closer to a continuous-wave, conventional diathermy machine. However, neither is more effective at providing heat therapy than a conventional diathermy machine, and each is more complicated and more expensive because they provide pulsed energy, a capacity deemed medically irrelevant by the FDA. Since, as the government conceded at argument, no one would purchase either machine solely for its heat producing qualities, the differences in pulse lengths seem to us irrelevant.

The FDA also argues that its approval of the Magnatherm was based on pre-1974 diathermy standards and that it has applied both pre-1974 and post-1974 standards evenhandedly to all to whom they apply. We might be sympathetic to this argument if the Magnatherm were an ordinary continuous-wave diathermy machine albeit one incapable of vigorous heating. However, the commercial value of the Magnatherm lies solely in its capacity to function as a Diapulse-like device, and we are unwilling to engage in a comparison of the minor differences between the two machines viewed as conventional diathermy devices. Such a comparison is a transparently disingenuous method of evading the realities of the issues before us, as the FDA fully acknowledged when it asked Judge Dooling to modify the 1974 injunction so as to bar DCA from marketing any device using a pulsed field. If the FDA was concerned solely about compliance with diathermy standards, condition (i) of the modified injunction was unnecessary. In fact, the FDA's real goal was to bar DCA from marketing anything but a conventional, continuous-wave diathermy machine, a goal obviously at odds with UME's marketing of the Magnatherm.

Moreover, comparing the P/emf and Magnatherm solely as diathermy machines leads directly to a result inconsistent with the FDA's regulatory mission. First, since neither machine is useful as a commercial matter solely as a diathermy machine, the FDA's decision is irrelevant in that regard. Second, if the FDA's scientific rejection of athermal electric healing effects is correct, it has not protected the public from the worthless devices by allowing the Magnatherm to be marketed. Third, if its rejection of athermal electric effects is incorrect, it has limited competition in the sale of medically worthy machines by outlawing the P/emf and Diapulse. This wholly anomalous situation cannot be allowed to persist.

We decline, however, DCA's invitation on its cross-appeal to second guess the FDA's rejection of athermal electric effects as a healing method. Evaluation of the relevant medical studies is better done by the FDA, and its rejection of the studies proffered by DCA is binding on us unless arbitrary, capricious, or an abuse of discretion. 5 U.S.C. Sec. 706(2)(A) (1982). The FDA's own studies of the Diapulse machine concluded that any therapeutic effects it may produce are solely a function of the slight heat it creates, not of the athermal electric effects. The FDA was entitled to attach greater weight to these studies than to DCA's submissions, most of which the FDA found "seriously defective in one or more respects."

However, we must insist that the FDA apply its scientific conclusions evenhandedly and that it not "grant to one person the right to do that which it denies to another similarly situated," Marco Sales Co. v. FTC, 453 F.2d 1, 7 (2d Cir.1971). So long as the FDA allows UME to circumvent its rejection of athermal electric effects, we have no choice but to allow DCA also to circumvent it. Deference to administrative discretion or expertise is not a license to a regulatory agency to treat like cases differently. We therefore affirm Judge Mishler on the appeal and on the cross-appeal.

We emphasize, however, that our decision leaves the FDA with considerable freedom. It may ultimately decide to cease treating both the modified Magnatherm and the P/emf as diathermy machines and remove both from the market; it may apply universally its curious policy of permitting a device to say it performs one function in order to be sold solely to do another; or it may heed the advocates of athermal electric effects and allow all Diapulse-like machines to be marketed. The FDA may not, however, grant to UME what it denies to DCA.

In view of the result we reach, we address DCA's claims of FDA bias only briefly. Henceforth, the FDA must apply to DCA the same scientific and legal standards it applies to DCA's competitors. We believe further relief is unnecessary since the showing of bias, if in fact bias has been shown, in no way taints the FDA's standing rejection of athermal electric effects, and does not, therefore, justify our undertaking to resolve that scientific dispute.

Affirmed.