In the world of FDA, there is a stark divide between the regulatory treatment of drugs and medical devices. A product intended to diagnose or treat disease, or to alter the structure or function of the body, is within the definition of a drug under the Federal Food, Drug, and Cosmetic Act (FDCA). If, however, the product has these intended objectives but does not achieve its primary intended purposes via chemical action on or within the body, or via metabolization, then it meets the statutory definition of a device.
The drug/device distinction drives enormous differences in premarket review and postmarket compliance requirements under the FDCA and implementing regulations. The differences are so significant that FDA long ago established separate bureaucratic centers devoted to each category — the Center for Drug Evaluation and Research (CDER) to regulate drugs and the Center for Devices and Radiological Health (CDRH) to regulate devices. Each center administers different statutory provisions (for the most part) and separate sets of implementing regulations. No one would dispute that a product regulated by CDER must meet very different requirements than one regulated by CDRH. No one would dispute, either, that user fees charged by FDA for drugs are far higher than for medical devices. For instance, in fiscal year 2019, a New Drug Application (NDA) for a prescription drug (with clinical data) has a user fee of $2,588,478, while a 510(k) submission for a device has a user fee of $10,953.
REGULATING DEVICES AS DRUGS
Some may be surprised to learn, therefore, that FDA occasionally regulates devices under the drug authorities. We are not speaking of products where it is a close call scientifically whether it meets the definition of a device. Nor are we concerned with combination product that may consist of both a drug and a device. Rather, we are addressing products that unequivocally meet the definition of a device, but FDA regulates them as drugs anyway.
An example is barium sulfate intended for use as a radiological contrast agent. Barium sulfate is an inert metal salt that absorbs X-rays to improve visualization of the gastrointestinal tract. It passes through the gastrointestinal tract without being absorbed or altered. All would agree – including FDA – with the fundamental science which says that the contrast agent function is not performed via either chemical action or metabolization.
Therefore, barium sulfate as a radiological contrast agent meets the definition of a device. For decades, FDA regulated these products as medical devices, granting 510(k) clearance to several products. Then, in a dramatic shift, FDA began regulating them as drugs. There are now numerous barium sulfate products for use as radiological contrast agents that have NDA approval instead of 510(k) clearance. A warning letter issued just last year said that a barium sulfate product is required to have approval under the drug authorities.
Another example is over‑the‑counter (OTC) skin protectants. These products are regulated as drugs under a final monograph (codified at 21 C.F.R. Part 347). Many of the permitted active ingredients and indications appear to meet the definition of a device. For example, consider the monograph-approved use of petrolatum for the permitted indication of preventing chapped or cracked lips due to the drying effects of wind and cold weather. It seems fairly obvious that this use of petrolatum does not rely upon chemical action or metabolization. Rather, the petrolatum serves as a barrier that covers the skin against the wind and helps it retain moisture. That is a classic device function.
FDA’S CORE ARGUMENT
There is plenty of administrative history and lore explaining how FDA assigned barium sulfate and skin protectants to CDER. But the purpose of citing these examples is to spotlight FDA’s general position. It is the agency’s assertion that any medical device can be regulated under the drug authorities for reasons of administrative policy and convenience. Yet, it seems clear that Congress has drawn a bright dividing line between drugs and devices, and has decreed that they are to be regulated differently. On what basis does FDA believe it is authorized to ignore the distinction and regulate devices as if they were drugs? FDA may only take such actions authorized by Congress, so the agency must establish that it has authority under the FDCA to regulate devices as drugs.
FDA’s core argument is based on the statutory definition of drugs and devices. FDA observes that the definitions, read literally, are not mutually exclusive. Rather, the drug definition encompasses any “article” intended to diagnose / treat disease or alter the structure or function of body. A device is an “article” for the same intended use, per the first part of the device definition. A device therefore meets the definition of a drug. It is true, FDA would agree, that the device definition has an added requirement that it may not utilize chemical action or metabolization to achieve its primary intended purposes. But, that simply means that not all drugs can be devices. According to FDA, it does not prevent all devices from being drugs. When it comes to devices, FDA says, the drug and device definitions overlap and FDA therefore has full discretion to classify a device as a drug. It is, therefore, by the grace of the agency that any particular medical device is not regulated as a drug.
IS FDA RIGHT ABOUT THE FDCA?
FDA’s “overlap” argument arguably comports with the literal words of the drug and device definitions. But, once one steps back and looks at the big picture, FDA’s argument is not supported by the basic scheme and operation of the FDCA, and is even contradictory to it. Let us look at the problems with FDA’s argument:
The dog that did not bark. Start with the fact that there is no provision in the FDCA expressly authorizing FDA to regulate products meeting the definition of a device as drugs. Given the lengths to which Congress went to set up completely separate regulatory schemes for devices and drugs, it is counterintuitive, to say the least, that Congress did not signal a grant of authority to FDA to regulate devices as drugs when it chooses to do so. If this authority truly had been granted to FDA, one would have expected Congress to set some parameters around it.
By way of contrast, there is a statutory provision allowing FDA to choose a lead center for combination products (FDCA § 503(g)). If a product combines a device and a drug, FDA is authorized to choose whether to regulate it in CDER or CDRH. But Congress did not leave the choice to FDA’s unfettered discretion. Rather, Congress requires that FDA make the decision based upon a determination of the “primary mode of action” of a combination product. In contrast, according to FDA, the agency has complete discretion for the equally significant decision to regulate a device in CDER. Why is there no analogous provision specifying the boundaries of FDA’s discretion in this regard? The most straightforward answer is that Congress did not grant such discretion in the first place and so had no occasion to specify the factors to be considered in making the decision.
“Congress does not hide elephants in mouseholes.” In FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120 (2000), FDA had claimed authority under the FDCA to regulate cigarettes and smokeless tobacco, on the ground that nicotine was a drug and cigarettes and smokeless tobacco were drug delivery devices. The Court did not dispute these products literally fit within the FDCA’s definitions of drugs and devices. Nonetheless, the Court concluded that FDA did not have the claimed authority. The Court declared “we are confident that Congress could not have intended to delegate a decision of such economic and political significance to an agency in so cryptic a fashion.” Id. at 160. A year later, in a different case, the Supreme Court put the same thought this way: “Congress … does not alter the fundamental details of a regulatory scheme in vague terms or ancillary provisions — it does not, one might say, hide elephants in mouseholes.” Whitman v. American Trucking Association, 531 U.S. 457, 468 (2001).
Here, once again, FDA has found an elephant hiding in a mousehole. It finds that a partial overlap in statutory definitions supports the rather spectacular authority to regulate all devices as if they were drugs. This authority has wide implications for the pathway to market for devices, the attractiveness of investment in device technology, and the cost and scope of life‑cycle regulation of devices. Surely, Congress would have said something somewhere in the FDCA about the exercise of this momentous decision‑making authority if FDA were intended to have it. Once again, the rational conclusion is that Congress did not grant FDA this authority and so did not mention it anywhere in the FDCA.
FDA’s position contradicts the plain language of Section 563. This provision is focused on the very topic under discussion, the classification of products as drugs or devices. It requires FDA, upon request, to uniquely classify each product “as a drug, biological product, device, or a combination product.” (Emphasis added.) It is plain from this language that Congress envisions that every product FDA regulates is uniquely either a drug, biological product, device, or a combination product. There is no option for classifying a product as a device “and” a drug. Yet, FDA claims that every device is simultaneously a drug. FDA’s position is contradictory to plain language of Section 563.
Section 563 also provides that a classification decision is binding on the agency and cannot be altered “except with the written consent of the person, or for public health reasons based on scientific evidence.” That allows a person to reliably know whether the product will be regulated as drug or device (or something else), with all the attendant differences between the two regimes. Yet, FDA’s position implies that the agency is authorized to override a Section 563 decision classifying a product as a device and to instead regulate it as a drug. If FDA may disregard a Section 563 classification decision without written consent, it obliterates the protection of Section 563. Hence, FDA’s position contradicts this aspect of Section 563 as well.
FDA’s position renders the device definition optional. A basic interpretative canon of statutory construction holds that “[a] statute should be construed so that effect is given to all its provisions, so that no part will be inoperative or superfluous, void or insignificant.” Corley v. United States, 556 U.S. 303, 314 (2009) (internal citations omitted). FDA’s theory of the overlapping definitions violates this canon. Under FDA’s approach, when a product meets the specific definition of a device, the agency may still regulate it as a drug. A decision to regulate a product meeting the definition of a device as if it were a drug renders the device definition inoperative in that case.
More generally, FDA’s position renders the device definition superfluous in all cases, by treating it as merely a suggestion from Congress, a starting point for decision‑making, while retaining for the agency discretion to disregard it. This construction of the statute should be rejected in favor of one that gives full meaning and effect to the device definition. Congress has made it clear that a product that does not utilize chemical action or metabolization to realize its primary intended purposes is to be regulated very differently than one that does. FDA does not have authority to breach this dividing line with a contrary decision.
FDA’s reliance on Bracco Diagnostics v. Shalala is misplaced. Finally, FDA sometimes supports its position by quoting Bracco Diagnostics v. Shalala, 963 F. Supp. 20, 28 (D. D.C. 1997). FDA always quotes the district court’s statement that: “The MBI products and plaintiffs’ products all likely meet both the definition of a drug and the definition of a device under the Federal Food, Drug and Cosmetic Act, and the FDA therefore has discretion in determining how to treat them.” Id. at 28.
FDA never quotes the next sentence, which states the court’s authority: “See 21 U.S.C. § 353(g) (‘The Secretary shall designate a component of the Food and Drug Administration to regulate products that constitute a combination of a drug, device, or biological product.’)” The court’s reliance on Section 503(g), relating to the assignment of combination products to a lead center, suggests that the district court viewed the products in question as “likely” combination products, with both device and drug modes of action. In fact, the specific holding of the court in Bracco Diagnostics was that FDA was required to treat functionally similar combination products the same way – putting them all in either CDER or CDRH. The court was not called upon to rule whether FDA may regulate devices under the drug authorities. Even if the statement could be read to support FDA’s position, it would be dicta, i.e., an incidental expression of a judge’s opinion, not essential to the decision and not establishing precedent.
FDA’s position that it may regulate device as drugs is plainly not authorized under the FDCA. To the contrary, if a product is within the definition of a device, Congress has decreed that it must be regulated under the device authorities. That is so even when it would be administratively convenient to do otherwise. Only an amendment to the FDCA could give FDA the authority it claims. FDA should act promptly to bring itself into conformity with the statute.