These are my notes from class meeting 8 of Harvard Law School’s Food and Drug Law course, led by Prof. Peter Barton Hutt on January 12, 2017. Reading for today’s class meeting is pp. 775-834 and 996-1044 of Food and Drug Law 4th Ed..

Special categories

Radiopharmaceuticals such as positron emission topography (PET) ligands are a distinct regulatory category within FDA. They were not originally under FDA jurisdiction, and when FDA did try to regulate them, the courts ruled in favor of manufacturers in Syncor International Corporation v. Shalala (1997). With FDAMA in 1997, Congress granted FDA jurisdiction in this area. FDA’s most recent Guidance on the subejct is FDA Oversight of PET Drug Products: Questions and Answers from 2012.

Medical gases were long a subject of dispute, as manufacturers held that they were grandfathered in from pre-1962 under GRAS or GRASE categories, while FDA viewed them as unapproved. With FDASIA in 2012, Congress finally resolved the issue; the 2015 Guidance Certification Process for Designated Medical Gases now lists the permitted gases and their permitted uses, and outlines procedures for approval of new gases.

FDA has jurisdiction that would allow it to regulate homeopathic drugs and Traditional Chinese Medicine products, but it has chosen not to regulate them.

Off-label prescribing

FDA has changed its position on off-label use over the years. For a while it held that unapproved uses of approved drugs were prohibited, but eventually it backed down. Following Congressional hearings over the off-label use of methotrexate for psoriasis in 1971, it became FDA policy that, for the most part, doctors have latitude over off-label prescribing; manufacturers are prohibited from promoting off-label uses, but can respond with information on off-label uses if specifically requested by a physician. There are exceptions, for example, any and all off-label use of human growth hormone (hGH) is strictly prohibited. In spite of the overall FDA policy, FDA has sometimes tried to enforce against physicians in specific instances of off-label use, but the courts ruled against FDA in United States v. Evers (1981).

Generic drugs

The Drug Price Competition and Patent Term Act (better known as Hatch-Waxman Act) of 1984 largely defined the way generic drugs are regulated today and led to a boon in generic production. Note that the opposite of generic drug is pioneer drug. The Act was a compromise between the interests of pioneer and generic manufacturers and provided a predictable regulatory framework. Pioneer drug manufacturers got a patent extension: half of the time spent in the IND stage and all of the time that FDA spends reviewing the NDA are added to the end of the patent term. Once that extended period is over, generic manufacturers can submit an abbreviated NDA (ANDA) that contains manufacturing information and enough human data to establish bioequivalence with the pioneer drug, but obviously does not need to contain whole new clinical trial data showing safety and efficacy. The usual bioequivalence data requirement is to dose 24 to 36 healthy volunteers just once and measure pharmacokinetic parameters such as Cmax and AUC, and bioequivalence data makes up most of the ANDA document. Hatch-Waxman also established the so-called “research exemption”, which makes it legal for people or companies other than the patent holder to manufacture and use an on-patent drug for research purposes. This made it possible for generic manufacturers to begin producing their drug, establishing manufacturing and quality control information, and doing bioequivalence studies before the patent term was up. FDA established the Orange Book, which lists all approved pioneer drugs and the patents that the pioneer manufacturer asserts (FDA does not review these assertions) apply to its drug. When a generic manufacturer submits an ANDA, it must, for each asserted patent in the Orange Book, explain whether it believes the patent I) was never filed, II) has expired, III) will expire, or IV) is invalid or irrelevant to the product at hand. The first generic (“first flier”) to be approved gets a 180-day exclusivity before any other generics can be approved.

The original implementation of Hatch-Waxman had an unexpected consequence. Often, the pioneer manufacturer would sue the generic manufacturer for patent infringement, and they would settle with a so-called “reverse payment”. Usually in patent infringement, the infringer has to pay the patent holder, but in these cases, the patent holder would pay the alleged infringer (hence “reverse”), and the alleged infringer would agree not to sell the drug. What’s more, if the infringer was the “first flier”, then by never starting to sell their drug, they could prevent the 180-day exclusivity clock from starting. Thus, by paying off the “first flier”, a pioneer drug manufacturer could effectively stall introduction of all generics. In 2003, Congress amended the law so that the 180-day exclusivity is forfeited if the first flier fails to actually introduce the drug into the market. A divided Supreme Court ruled in 2013 that the FTC can also go after these deals as being anti-competitive, although the burden of proof is on FTC, as reverse payments are not “presumptively unlawful”. (Dissenters felt that the reverse payment deals were fine).

Hatch-Waxman does not apply to biologicals; there is a separate procedure for generic “biosimilars”.

Drug prices

FDA published a 2006 study of Generic Competition and Drug Prices. They found that when there is only one generic product competing with the pioneer drug, the price only drops by 6%. Two generic products drops the price by about half, and as you add more and more generics in competition with each other the price eventually drops by more than 90%.

Drug pricing is very controversial. Multiple observers, including Malcolm Gladwell, have pointed out that compared to the rest of the developed world, the U.S. has the highest on-patent drug prices but the lowest generic drug prices [Danzon & Furukawa 2003]. He also pointed out that (at least as of 2004) a lot of the year-on-year increase in total drug expenditure was due to increases in the number of prescriptions, and only a fraction was due to increases in prices. There are several proposed and attempted responses to drug pricing, including allowing importation from Canada (discussed in lecture 2), and having Medicare or states either promote or require the use of generics where available, set limits on payments, or negotiate prices.

NewCo, Part III

This is the third in a series of three episodes of a narrative about a fictional biotech startup company, NewCo, developing a new drug.

Data from the Phase I trial are in, and in addition to an absence of any safety issues, there is a clear dose-response, where the 10 and 20 mg doses perform poorly, the 30 mg dose is about as good as the average optical colonoscopy, and the 50 mg dose provides superior quality, among the best ever seen. The next question is whether NewCo should proceed to a Phase II to directly to a Phase III, and whether it wants to stick with the 50 mg dose or test additional, even higher, doses. The main consideration in Phase II vs. III is that a Phase III is much more expensive, so it is a bit of a risk to jump straight to it without first confirming efficacy in a Phase II — after all, the efficacy data from Phase IB are considered preliminary. Also, the Phase IB was performed using the natural extract, whereas (it’s now two years later) we hope we are close to isolating and synthesizing the active ingredient. The Phase IB data might be just barely enough to earn a CE Mark in the E.U., but because efficacy in colon cancer hasn’t actually been demonstrated, no doctors would actually prescribe it based on the data available today, so that wouldn’t be particularly profitable. European physicians tend to be more conservative in that they do not switch to new products until superiority has been more conclusively demonstrated, whereas U.S. physicians are more willing to be early adopters. Plus it would be a CE Mark on the natural extract, whereas what we ultimately want is approval of the isolated (and more patentable) active ingredient. The Phase II would cost $4 million, allow us to confirm the right dose, and take 1 year during which the company would burn another $6 million. It would put the company in a better negotiating position with VCs, where it could raise the money for a Phase III once it’s got more efficacy data. And the Phase II trial data could be published, thus starting to make doctors aware of the new technology and its promise.

The company decides on a Phase III. At this point, the isolation and synthesis is still not complete, so the Phase III will use the natural plant extract. The trial may be able to have multiple arms with an even higher dose such as 70 mg, since it has seen no evidence at all of toxicity, though this will have to be negotiated with FDA. Although the Australian Phase I worked out great, NewCo wants to do the Phase III in the U.S. because it perceives that U.S. physicians will be more likely to actually use this product after approval if they see that its efficacy has been demonstrated in the U.S. context. NewCo needs to have two adequately controlled trials, so sit uses the “Mississippi strategy” — it splits the U.S. down the Mississippi River and defines one trial as being west of the Mississippi and the other as east. In total, it wants to enroll 2,000 people, and since this is a U.S. trial, it expects to spend $30,000 per participant. Therefore the costs of the trial alone are $60 million, plus the company needs money to keep its doors open while the trial is running. The trial will use two surrogate endpoints: image clarity (not validated, but FDA will want to see the data anyway), and detection of small polyps (considered a validated endpoint meriting full approval). Note that a true clinical endpoint would be something like death due to colon cancer, which would take decades and would clearly be infeasible. The study design will be crossover — each participant will undergo both a virtual and an optical colonoscopy. The goal is to show superiority over the optical approach. To try to enrich for people with high risk (and thus power to show improved detection of polyps), they may try to specifically recruit people with a family history of colon cancer. The main target population is past reproductive age, so they probably don’t need to design the trial so as to be able to detect the presence or absence of reproductive toxicity, though it might be something to be monitored post-approval. Typical followup time will be 6 or 12 months. A longer (perhaps 5 year) follow-up might be required if there was specific reason for concern about carcinogenicity, but because there is no indication of genotoxicity from preclinical studies and the dosing is very occasional, FDA probably won’t require this.

We also return to the issue of incidental findings. The fact that full body scans are specifically recommended-against by all major medical bodies means that we should not tout our product as having the ability to find cancers in other organs, and in fact, the procedure should be to do the CT scan only over the range of the bowel. The scan should stop at the bottom of the lungs and at the pelvis, thus reducing radiation. (There is no need for lead shielding as the CT scan radiation is confined to a ring of exposure and the technician can determine how far along the torso to scan).

To do anything, though, the company needs to raise more money. There ensue hardline negotiations between management and existing investors about the price per share that the investors should pay at this stage. They bought their original stakes for $1/share; based on the new data, management thinks new shares should be issued at $5/share, while the investors downplay the results, saying $2/share would be reasonable.

The phone rings: the IRB for the academic hospital at University of Buffalo has rejected our IRB protocol for the Phase III, saying they believe it is unethical to proceed directly to Phase III without doing a Phase II first. NewCo decides to shop around for a different IRB that will approve its original Phase III plan. This happens frequently and FDA does not take a negative view of companies going to one IRB after another until they get their protocol approved. The philosophy is that each IRB is supposed to reflect local ethical standards and expectations, rather than any universal standard. Moreover, NewCo is only going to shop around to other very well-respected medical centers, which no one could argue represent a low standard of ethical conduct.

The Phase III trial data come in. There were no significant adverse events. Surprisingly, the 50 mg dose, which looked so promising in the Phase IB, provided inferior image quality, and detected nominally (though not significantly) fewer polyps than an optical colonoscopy. Luckily, the company also tested a higher, 70 mg dose, which gave above-average image quality, and nominally (though not significantly) more polyps than an optical colonoscopy. Based on this, the company can’t demonstrate superiority, but it can ask FDA for approval on the basis of equal efficacy combined with better tolerability.

NewCo prepares an NDA and decides to request Priority Review, meaning 6 months instead of 12 months. In view of the millions of dollars per year that NewCo expects to make after approval, the 6 month difference is an enormous financial value. The company has by now developed a total synthesis of the active ingredient, and is already budgeting money for a bridging study to be carried out after the NDA approval. FDA is convening an Advisory Committee on NewCo’s product, and NewCo is preparing for it with greater vigor than a lawyer would prepare for a Supreme Court case. The meetings are public and always attended by the press, NewCo’s stock price will yo-yo up and down during the hearing in real-time in response to press coverage of every comment, and if the committee recommends against approval, the stock will probably lose 75% of its value. Advisory Committees only make recommendations, but FDA follows the committee’s recommendation about 80% of the time. A major decision is whether company employees will make the presentation, or whether the company will invite a world leading physician-scientist in colonoscopies to make the presentation. Whether the committee will take one or the other more seriously is pure speculation. The outside scientist would need to not have stock options, but would of course be paid an hourly consulting fee, and this is not considered to preclude them from speaking or being credible. That said, everyone in the room will be asked whether they are being paid to be there, and the voice of patients and their advocates who have flown in on their own expense are extremely powerful.

FDA reviews NewCo’s raw data, and finds one anomaly: one of the several study site investigators has failed to collect or file all of the required documentation. There is no evidence of fraud but the data are considered incomplete, and the investigator is off the grid on vacation. NewCo has no choice but to remove that site’s data from the NDA. Luckily, the removal of this subset of data doesn’t meaningfully change the statistics. Meanwhile, FDA mentions it would like to discuss labeling. This is an incredibly good sign: FDA only ever wants to discuss labeling if it is very likely to approve the drug. Because of the anomalous animal finding about litter size, FDA wants to play it safe and create a registry of pregnant women getting the drug. This is doable, but doctors hate having to enroll patients in a registry, so this would dramatically reduce the market size for our product. Our other option is to go to dispute resolution with FDA. The informal approach is to simply ask for a meeting with the next highest official (above the person who initially made the decision), and if necessary, keep appealing upward. Alternately there is a formal dispute resolution process created by the PDUFA legislation. NewCo’s VP of Regulatory Affairs, conservative as is typical for his position, wants to just comply with FDA and not seek dispute resolution, because he worries about offending FDA officials and poisoning the well for future interactions. One option is to request a meeting with FDA and discuss other options — perhaps a statement in the label about the anomalous result in the animal studies, or even pregnancy warning on the label, or at most an outright contraindication for pregnancy. One downside of the latter two options is that once you get a restriction or warning on a label, it is extraordinarily difficult to ever get it removed.

As much as NewCo’s luck is looking great, with the drug verging on approval, it is running on fumes in terms of budget: its flameout date is only 2 months away. If it wants to commercialize the drug itself it would need to raise $150 million and hire a ton of people to do marketing and sales and so on, but that’s a large mountain to climb: more likely, NewCo will partner with a major pharma that will handle all the marketing and distribution and so on. By this point, NewCo should long since have had offers from potential pharma partners; if it doesn’t, that means something is gravely wrong. Meanwhile, the original VCs are thinking about an exit: they’ll either 1) need new investors, 2) sell the whole company to a major pharma, or 3) make an Initial Public Offering. Options 2 or 3 are most likely. But either negotiating a merger & acquisition, or planning an IPO, will take several months. The CEO finally decides to ask existing investors for a bridging loan just to keep the company running long enough to get to an IPO.

The NDA gets approved, and NewCo has to decide about pricing. The competing virtual colonoscopy product (which is considered inferior to optical) costs $800 per use, and a typical optical colonoscopy costs $1500 per procedure. Because NewCo’s product has been approved on the basis of superiority over optical colonoscopy, the CEO feels it can price it several fold higher. To get the best return without too many eye-popping zeroes that attract public and Congressional outcry, management considers a price of $9,900 per use. About 50% of the market will be covered by Medicare, which cannot negotiate prices, so they think the market will bear this. However, a board member speaks up to strongly argue that the price should be $1500 or less, because third party payors and to some extent Medicare will find ways to not pay for this if it costs $9900, and therefore market will be small, which will mean it takes more years to gather enough data to prove that this product has a mortality and morbidity benefit in the long run. At $9900, the product will remain a niche, small market for the duration of its patent, whereas at a competitive price like $1400, it will quickly sweep a large fraction of the market, and within years replace optical colonoscopy as standard of care. The CEO finally settles on a list price of $2500, with a complicated program of discounts to extract every customer’s maximum willingness to pay, thus maximizing the market size while earning an average actual revenue of about $2000 per use. The IPO takes off at $20/share, meaning the early VCs made a 20:1 return.

the end