410 Century at FDA Robert J Temple M

4/10 Century at FDA Robert J. Temple, M. D. Deputy Center Director for Clinical Science Center for Drug Evaluation and Research U. S. Food and Drug Administration Dr. Harvey Wiley Lecture April 23, 2014

It’s a Great Place I won’t say we never worry about externalities (politics or press), but it doesn’t (well, hardly ever) affect what we do. What I have seen and experienced is an organization that single mindedly and more than everything else wants to get the right answer. The right answer can change; how to get it and analyze it can change, and people can of course disagree, but the goal is shared and strongly held. And the people are wonderful from Dick Crout and Marion Finkel when I got there, to Janet Woodcock now, and Bob O’Neill pretty much throughout (and too many others to name). 2

Uniquely Placed Given our role and authority we are, of course uniquely influential. A good idea at FDA spreads like no one else's; so its not that we invent or think of everything, but when we take up an idea we can make a huge difference. I’ll illustrate this with some of my trial design obsessions, then note a few other areas. But you will see, I hope, that we have greatly influenced the improvement over the years in clinical trial design and conduct. So here’s what I feel good about. 3

A & WC Studies The world changed in 1962, with the brilliant definition of substantial evidence as arising ONLY from adequate and well-controlled studies, but frankly, we did not know too much about adequate and well-controlled studies, first attempting to describe them in 1970. Mind you, some people knew. Lou Lasagna knew, Paul Meier and some people at NIH and the VA knew and a clinical pharmacologist whose statement of principles Henry Simmons cited in the late 1960 s, Bill Beaver knew, but we were just starting. But we started in the 70’s and continued since then and the improved standards have long been in place. The studies used to support NDA approval are almost always 4

Learning So how did we learn? 1. DESI We had to review 100’s of studies, done every imaginable way, but we learned plenty. Analgesic case: subsets post-facto 2. Accumulating experience in cardio-renal (1978 -1980) • Nadolol for angina: insight into NI. • Chlorthalidone dose-response - the randomized fixed-dose-response study. Also excess dose of nadolol. • Cimetidine – stopping rules • Anturane Reinfarction Trial: CV death analysis, counting all patients 5

So What Followed 1. DIA Workshop - 1981 First suggested “equivalence” studies need a revisit; with wording and recommendations resembling the NI guidance. This was written up in the DIA journal in June 1982. NI guidance proposal (2010) began then, 29 years ago. 1983 (final 1987) A& WC revision of 314. 126 clearly understood NI problem and said: “If the intent of the trial is to show similarity of the test and control drugs, the report of the study should assess the ability of the study to have detected a difference between treatments. The analysis of the study should explain why the drugs should be considered effective in the study. ” • ICH E-10, 2000 discussed NI studies at length, making similar points. • NI guidance (draft) in 2010. • Now broadly understood, a “sea change, ” but began with nadolol, the “light” 6 switch.

So What Followed 2. DIA – other matters suggested in 1981 -2 • Suggested a study now in our enrichment guidance; studies in non-responders, randomizing to new and failed drug to show an advantage in the non-responders population. Used for captopril approval. • Suggested abandoning titration and moving to randomized, fixed-dose, dose-response studies. 7

Dose-Response n The chlorthalidone D/R study changed everything. All studies had used titration designs, which give no interpretable D/R. Impressed by evidence that chlorthalidone was used in 4 X excess and probably caused fatal arrhythmias, data coming from 2 randomized, fixed-dose, dose-response studies, we began asking for those studies and began seeing them regularly, with a huge improvement in dosing. n ICH E-4 (Dose-Response Information to Support Drug Registration, 1994) recommended that design, now the norm. 8

Dose-Response 3. Enrichment In the early 1980’s I gave a lecture in an NIH course on clinical studies organized by Dave De. Mets, Larry Friedman, and Curt Furburg. I spoke about “ENRICHMENT” study designs, a term never used. The concept evolved, led to a publication by me and publications by others, and (finally) to our draft guidance in 2010. As in many cases, enrichment maneuvers were regularly used but the concept needed focus, and we helped provide it. 9

What Else? 4. Counting all Patients The Anturane Reinfarction Trial was a (sort of) model, the first outcome trial (I think) by industry (Ciba-Geigy), with a distinguished steering committee, etc. But not so great on endpoints, pre-specification of analyses, and counting all people, and we learned a lot. It was a CV mortality study in patients 25 -35 d post MI • Originally all deaths, then divided CV into MI, sudden and other, and looked at deaths overall and at 6 months, etc. , but with NO PLAN for all this. • Did best on 6 month Sudden Death, but this classification was bogus – same event called SD on placebo, MI or other on A, giving a seeming advantage of A on SD. • Even worse completing patients WHO DIED were excluded for poor compliance or inappropriate entry, 8 A deaths, 1 placebo. Put them back, no effect at all. These exclusions were not planned and not 10

Counting All Patient ART taught us a LOT • Clin Stat Guidance (1988) RT and O’N, makes it very clear that all patients must be accounted for. Any plans to exclude patients with data should be presented. • ICH E-3 (1996) [Structure and Content of Clinical Study Reports] also makes it clear that all patients must be accounted for and any patients not assessed should be explained. 11

Other Stuff 1. Demographics Early 1980’s growing interest in elderly. NIA asked us to do a guidance, and we drafted it in 1983, finalized it in 1989 and an ICH guidance (E-7) was written in 1994, with an update in 2012 to (finally) recognize that you need people over 75. Guidance on including both genders was proposed in 1993. In 1985 the NDA regulations hinted at an interest in differences of S and E in subsets, but in 1998 regulations plainly required analyses of S and E (in ISS and ISE) by demographic group. The Clin-Stat guidance (1988) and ICH E-3 had already strongly called for these. FDASIA report shows these analyses are being done regularly. 12

Other Stuff 2. Access Hard to believe, but the Treatment IND concept was proposed in 1983, final in 1987. In fact, this had been done in 1970’s (metoprolol, nifedipine) when new treatments were late in development and represented obvious advances. Recently growing interest in early availability of breakthrough types (cancer, virus and other infectious disease, genetic disorders), where very early data can be convincing. We have generally been making them available. 13

Other Stuff 3. Help and Communication Great interest in FDASIA breakthrough provision, but the concept of special attention to drugs for serious illness with no good Rx has been around for a long time. • • The 1983 proposed IND rewrite (which proposed the Treatment IND), specifically notes the importance of FDAindustry meetings to “facilitate the drug development and approval process” and particularly noted EOP 2 meetings and tried to facilitate having them. Subpart E (1988), directed at drugs for life-threatening disease without satisfactory treatment set forth the need for extensive early consultation, flexibility, accepting greater risk. Fast-track and now breakthrough capture this attitude and 14

Other Stuff 4. Accelerated Approval AA was another response to urgency, proposed and completed in 1992. It was a brilliant “compromise” on what needed to be shown for approval, but not how, reflecting the importance of responding to urgent needs and still getting the evidence that you really are helping. Huge difference in oncology, viral illness, genetic disorders. 15

Safety I’m not sure it’s appreciated, but the two most common single causes of drug withdrawal, hepatotoxicity and QT-interval prolongationinduced arrhythmias, will rarely if ever, happen again because we anticipate liver toxicity (Hy’s Law) and have spared the US three hepatotoxicities marketed (then withdrawn) in Europe, and we anticipate QT-prolongation with the uniformly conducted thorough QT study. A clear pair of safety steps forward. 16

Bottom Line FDA presents a never-ending array of fascinating issues – clinical, statistical, tactical, philosophical and the solutions are complex, fascinating, variable, and critical to the development of effective and safe drugs. These include: • • Etc. Trial design, novel approaches (enrichment, adaptive) Weighing benefit and risk New kinds of analyses and what to make of them Explaining what you did and why 17