FOOD AND DRUG ADMINISTRATION

P-R-O-C-E-E-D-I-N-G-S

(7:34 a.m.)

DR. NELSON: Good morning. Welcome to the meeting of the Pediatric Advisory Committee. I'm Dr. Nelson, the Chair, and I'll turn it over to Jan to start our business. We'll do introductions first.

As a quick reminder, you have to push the mic on and off. Let's start with introductions. We can start down at this end, and keep them brief.

DR. GAROFALO: Yes. I'm Dr. Elizabeth Garofalo. I'm the Industry Representative, and I am an independent consultant.

DR. GORMAN: Dr. Richard Gorman, pediatrician in private practice, and the representative of Professional Healthcare Organizations.

DR. LESLIE: Laurel Leslie, Behavior and Developmental Pediatrician at Children's Hospital, and a Health Services Researcher.

DR. VITIELLO: Ben Vitiello, National Institutes on Mental Health.

MS. KNUDSON: Paula Knudson, Community Representative, IRB Administrator, University of Texas Health Science Center - Houston.

DR. WARD: Dr. Bob Ward, University of Utah, I'm a Neonatalogist and Pediatric Clinical Pharmacologist.

DR. PINE: Danny Pine. I'm a Child Psychiatrist from the National Center of Mental Health.

DR. O'FALLON: Judith O'Fallon, retired statistician from the Mayo Clinic.

MS. DOKKEN: Deborah Dokken, Patient Family Representative.

DR. DAUM: I'm Robert Daum, Pediatric Infectious Diseases from the University of Chicago.

DR. DIAZ: Angela Diaz, Director of Mount Sinai Adolescent Health Center in New York City.

DR. MOORE: John Moore, Pediatric Cardiologist, University of California - Los Angeles.

DR. NELSON: Robert Nelson, Pediatric Care, Children's Hospital in Philadelphia.

DR. JOHANNESSEN: Jan Johannessen, Executive Secretary, Pediatric Advisory Committee.

DR. RAPPLEY: Marsha Rappley, Developmental and Behavioral Pediatrics, Michigan State University.

DR. HUDSON: Melissa Hudson, Pediatric Hematology Oncology, St. Jude Children's Research Hospital.

DR. BIER: Dennis Bier, I'm a Pediatric Endocrinologist and Nutritionist from Baylor College of Medicine in Houston.

DR. NEWMAN: Tom Newman, Child Pediatrician and Professor of Epidemiology and Biostatistics in Pediatrics at UCSF.

DR. MURPHY: Diane Murphy, Pediatrician, Office of Pediatric Therapeutics, FDA.

DR. LAUGHREN: Tom Laughren, Division of Psychiatry Products at FDA.

DR. ANDREASON: Paul Andreason, Division of Psychiatry Products, FDA.

DR. DAL PAN: Gerald Dal Pan, Division of Drug Safety, FDA.

DR. ROBERTS: Rosemary Roberts, Pediatrician, Office of Counter-Terrorism and Pediatric Drug Development.

DR. IYASU: Solomon Iyasu, Division of Pediatric Drug Development at FDA.

DR. NELSON: Thank you, and I'll turn it over to Jan for the Conflict of Interest Statement and Introductions.

DR. JOHANNESSEN: Thank you and good morning. The following announcement addresses the issue of Conflict of Interest with regard to today's discussion of a report by the Agency on Adverse Event Reporting as mandated in Section 17 of the Best Pharmaceuticals For Children Act for Clofarabine, Irbesartan, Sibutramine, and the mixed salts Amphetamine Product, and two, discussion and reports on Neuropsychiatric and Cardiovascular adverse events possibly related to other approved ADHD medications. This statement is made part of the record to preclude even the appearance of such at this meeting.

Based on the submitted agenda for the meeting, and all financial interests reported by the Committee participants, it has been determined that all interested firms regulated by the Food and Drug Administration present no potential for an appearance of conflict of interest at this meeting. In the event that the discussions involve any other products or firms not already on the agenda for which an FDA participant has a financial interest, the participants are aware of the need to exclude themselves from such involvement, and their exclusion will be noted for the record.

We note that Dr. Jorge Armenteros, Dr. Laurel Leslie, Dr. Cynthia Pfeffer, Dr. Daniel Pine, Dr. Ben Vitiello, Dr. Robert Moore are participating in the meeting as Voting Consultants, and that Paula Knudson is participating as the Acting Voting Consumer Representative. We would also like to note that Dr. Elizabeth Garofalo has been invited to participate as a Non-Voting Industry Representative, acting on behalf of regulated industry. Dr. Garofalo is an Industry Consultant. Dr. Richard Gorman is participating as the Non-Voting Pediatric Health Organization Representative acting on behalf of the American Academy of Pediatrics. With respect to all other participants, we ask in the interest of fairness that they address any current or previous financial involvement with any firm whose product they may wish to comment upon. Thank you.

DR. NELSON: So while Solomon is making his way to the podium, let me just make a quick comment. You'll notice, members of the Committee, that we don't have a lot of time for questions. I'm going to ask that as these presentations are being done, you write down your questions on the assumption that maybe they'll be answered by the next speaker, rather than taking time after presentations, and then we'll hopefully get done what we need to get done in an efficient manner. And I'm also going to ask the FDA folks to try and stay on time, as well, since we'll be expecting that from everybody. Solomon.

DR. IYASU: Good morning. I'm going to introduce the Adverse Event reporting under BPCA. This will be the subject of discussions for the first one hour, and then later, as well. As you know, we've done several of these presentations in the past. Almost 50 drugs have been reviewed under BPCA. Of course, this activity is mandated under Section 17 of the Best Pharmaceuticals for Children Act, which mandates that for drugs that have been given exclusivity, that during the one-year period after exclusivity, any reports of postmarketing adverse events would be reviewed and then presented to this Committee for their review and any recommendations.

Of course, most of the reports that we are mandated to discuss here come from the database of all adverse event reports that were received by the FDA. This has a long history which started in 1969; it has almost more than two million reports. This database contains hundreds of export drugs and therapeutic biologic products, with the exceptions of vaccines.

As you know, this is a spontaneous voluntary system. Most of the reports come from manufacturers, but there are also reports that come from healthcare professionals, consumers, and patients. More than 90 percent of all those reports that we have in the AER system come from manufacturers, because they're required under the law.

Just to give you some background, the definitions that we use for such reports is defined under the law. It does not require that the reporting needs to be -- that the adverse event needs to be causally related and proven before reporting. Any adverse event, whether or not considered drug-related, is to be reported to the AER system, so I think accidents, intentional overdose, any adverse events occurring from abuse or drug withdrawal calls for action.

Also, to just update the Committee, and for those who are new to this area, unexpected adverse drug event is also defined under the law, and it's defined as an event not listed in the current labeling for the drug product, including events that may be symptomatically or pathophysiological related to a labeled event, but differ because of greater severity or specificity.

Again, there is a regulatory definition, just to give you some background, that a serious adverse event is any event occurring at any dose that results in any of the following outcomes, and these outcomes are death or life-threatening adverse event, hospitalization, or prolonged additional hospitalization, persistent or significant disability or incapacity, any congenital anomaly or birth defects, or others that may require some intervention, so there is a regulatory definition for serious adverse event.

Of course, one has to assess these reports as they come, and there are many factors that effect whether there's a causal relationship between a report and the assessment. And key among these factors that we have to look at FDA are whether there is a temporal relationship between the event and the drug, that the event occurred after the drug is a strong indicator that there might be some relationship.

If we have, of course, information about those events regarding the challenge or re-challenge, meaning that the AER or the adverse event subsides when a drug is discontinued, or it reappears when a drug is re-administered, is also suggestive of a relationship. If we see a dose response in this report, that's also another suggestive information that might help us understand whether there's a causal relationship or not.

Again, we look at the biologic plausibility, looking at prior knowledge about the PK and the pharmacokinetic actions of these drugs. We also look at animal clinical studies, and any laboratory evidence that may be part of the report that we get.

Another factor that we look at is whether there is a known class effect, that a drug belongs to a class of drugs that has been proven to have a certain effect that we're looking at. And very important consideration has to be given also to underlying disease and concomitant drugs, because many of these reports that we get are highly confounded by underlying disease that may be exhibiting the same adverse event that we're looking at.

Now there are several strengths and also limitations to the AER's database, so when we look at postmarketing reports, one has to be very careful that it includes all drugs that are marketed in the United States, and, therefore, it's a very large sweep of what's going on. A simple, inexpensive system provides for mostly early detection of safety signals, especially if they are rare adverse events that are serious. For example, anaphylaxis, liver failure, or plastic anemia, and serious events like that.

There are also some important limitations. Under-reporting is one of the key limitations of the system because they are spontaneous reports. And the extent of under-reporting may vary from drug-to-drug and over time. The quality is variable, and because the numerator, which is events, uncertain in terms of numbers, the denominator also in terms of exposure is also uncertain, so often it's very difficult to quantify the risk of these drugs with respect to adverse event.

Now often the gold standard for looking at causality between an adverse event and a drug is the randomized control trials. They are, of course, the best for establishing causality and provide a non-biased estimate of the risk of a particular drug with respect to a suspect drug. So especially very important to have comparative data from these kind of studies to assess when the events under consideration have a high background rate. But, of course, RCTs also have their own limitations. They often are short in duration, and thus have a very small sample size and, therefore, they may not pick up rare adverse events. And often we pool safety data across many different randomized trials to improve the precision and power to detect such events. But, of course, there are limitations to doing that because the different RCTs that we pull together may have different methodologies that may limit our ability to detect some of the events.

And also, another important feature is that RCT's have this more or less select population, and are homogenous, and what you would expect to see in daily clinical practice; therefore, it drives the ability of the risk estimates that we get from RCTs may not apply to daily patient population. Therefore, there are some limitations even to RCTs. Of course, not finding an association from an RCT does not rule out the safety issue when you have a preponderance of reports, so it's a vicious cycle that you have to --if you don't get information that is confirmatory, and have you have a preponderance of reports, then you have to be concerned.

Therefore, today we'll continue our practice of presenting the adverse events under BPCA, and we'll provide you a review of the safety events during the year. And as appropriate, we'll also look prior to the one-year period that's mandated by BPCA to review the adverse events that are reported to FDA when it's warranted, and that we have some serious adverse events prior to the one-year.

We also provide you information on the drug use of the particular drug that we're looking at, first for out-patient, as well as for in-patient use. We also, for context, provide you a summary of the clinical, pharmacology, toxicology reviews of the exclusivity studies. We provided the drug label, published literature, responsive materials, and presentations, so that you have a comprehensive set of information to look at.

As we've done in the past, we'll have -- some drugs are going to be discussed in an abbreviated fashion when there's no new safety concern, some are going to be standard presentation when there are unlabeled new safety concerns, and then we'll have drugs that have in-depth presentation, because there is a safety concern that we'd like you to look at.

Today we have four drugs that are scheduled for BPCA safety review. Clofarabine and Irbesartan will have an abbreviated report, because we don't feel that we have concern regarding safety for these two drugs, just from the review that we've done. Sibutramine will have a standard presentation. We'd like to give you a little more, because there's been extensive reviews of cardiovascular, as well as fetal toxicity effects of these drugs and, therefore, we'll have two presentations on this so that we have a comprehensive history of what the state of knowledge is about Sibutramine regarding these two issues.

And then the last drug that will be discussed today on the schedule out of BPCA is Adderall, mixed amphetamine drug. It's providing an in-depth presentation of this as part of the post exclusivity -- as part of the session on the entire safety concern of all ADHD medications later today.

I'd like to acknowledge the many offices that have helped with all these reviews, and our first presentation, I'll just get right to it, will be on Clofarabine. Dr. Larry Grylack from the Pediatrics Division will present adverse event reviews for Clofarabine. Dr. Grylack has been a medical officer in the FDA for more than three years. Dr. Grylack.

DR. GRYLACK: Thank you, Dr. Iyasu. It's an honor to be here again this morning presenting to the Advisory Committee, and apologies to the residents of the Pacific and Mountain time zones for the early hour. This is an abbreviated presentation of the Adverse Event Review for Clofarabine. The drug is marketed as Clolar for intravenous injection. Its therapeutic category is anti-neoplastic in the sub-category of purine nucleoside anti-metabolite. The sponsor is Genzyme Corporation.

The indication in the Clolar label is the treatment of patients 1-21 years old with relapsed or refractory acute lymphoblastic leukemia after at least two prior regimens. Pediatric exclusivity was granted for Clolar in patients 1-21 years of age on July 14^th, 2004, which was prior to the marketing approval date of December 28^th, 2004. Of note is that Clolar did not have prior marketing approval for any age.

The in-patient, out-patient, and sales databases used by the FDA document no use of Clofarabine. Also, the Children's Oncology Group database reported no usage. The sponsor estimates 200-300 uses per year.

The FDA Office of Drug Safety's Adverse Event Search covered the period from the time of market approval, December 28^th, 2004, until August 14^th, 2005, which was the one-year post-exclusivity date. There were 12 unduplicated pediatric adverse event reports, eight of which were domestic reports. Of the 12 reports, four were deaths, and five were hospitalizations.

Details of the four deaths that occurred in patients receiving Clofarabine during the exclusivity review period are presented on this slide. One patient suffered a cardiac arrest after an episode of aspiration and treatment with airway suction occurring 48 hours after the fifth dose of Clofarabine. The patient had a history of cardiac dysfunction, and post-mortem exam revealed cardiac fiber hypertrophy.

The second death was associated with tumor lysis syndrome, resulting in multi-organ failure one day after the start of Clofarabine treatment. The third death was described as being caused by progressive disease one month after the diagnosis of capillary leak syndrome occurring within 24-hours of the first dose of Clofarabine. The fourth death reported was due to a cardiac arrest associated with an error in medication not being used for the treatment of the leukemia. This death occurred seven weeks after treatment with Clofarabine.

This completes the post-exclusivity adverse event reporting as mandated by the Best Pharmaceuticals for Children Act. Most adverse events are currently labeled or would not be unexpected in association with a disease, or with the concomitant treatments received by the patients. The Food and Drug Administration recommends routine monitoring of Clofarabine for adverse events in all populations. We would like to know whether the Pediatric Advisory Committee agrees.

Finally, I'd like to acknowledge all of these individuals from the Office of Drug Safety, Office of New Drugs, and the Office of Clinical Pharmacology Biopharmaceutics who provided reports which contributed to my presentation today. I would also like to acknowledge Denise Pica-Bronco who has admirably filled Kristin Fucas' role of coordinating our Advisory Committee presentations within the Division of Pediatric Drug Development. Thank you for your attention.

I would like to next introduce one of my colleagues, Dr. Alan Shapiro. Dr. Shapiro is a Pediatric Infectious Disease Specialist with a Ph.D. in Biochemistry. His past research includes work in immunology, infectious diseases, and molecular pharmacology. He has also had training in Pediatric Nephrology and Medical Genetics. Dr. Shapiro has been with the Division of Pediatric Drug Development for over two-years working as a medical officer. Really, the FDA should give him more than two-years credit considering all his many talents, training aspects. Dr. Shapiro.

DR. SHAPIRO: Thank you, Larry. Now I'd like to discuss the adverse events for Irbesartan. Irbesartan, also known as Avapro, is an anti-hypertensive in the Angiotensive II receptors class. Its sponsor is Sanofi-Synthelab. Its indication is for the treatment of hypertension alone or with other anti-hypertensive agents, also for the treatment of diabetic nephropathy in patients with type II diabetes and hypertension. Its original market approval was September 30^th, 1997, and exclusivity was granted on September 16^th of 2004.

To summarize, no pediatric adverse events were reported to AERS during the year following exclusivity. We did find a raw count of four adverse events, three serious, since market approval. When we did a hands-on review of these adverse events, we only found one unique adverse event report. That consisted of a patient with urticaria and erthema multiforme minor in a seven year old clinical trial participant who had a vial respiratory infection. Urticaria and hypersensitivity are described in the product label.

Okay. The use of Irbesartan in pediatrics is small, only about 3-4,000 prescriptions per year, and represents only 0.1 percent of the total prescriptions. Now going on to the clinical trials. In the dose ranging trial, pediatric patients with hypertension done for pediatric exclusivity, the sponsor failed to show a dose response relationship. Only a small treatment effect on blood pressure during the randomized placebo controlled withdrawal phase of the trial was observed.

Based on these studies, Irbesartan does not appear to be useful in the treatment of hypertension pediatric patients aged six years and above, despite receiving doses that are equivalent to a plasma concentration seen in adults.

Going on to the history of the pediatric exclusivity for Irbesartan, a written request was issued in 1998, and then we received a submission of pharmacokinetic data in February of 2000, which resulted in a labeling change in May of 2001.

Final submission of exclusivity studies in 2004 resulted in a proposed labeling revision in October of 2004 that would remove the pharmacokinetic data from the label because the pediatric indication was not approved. I should mention that this labeling change is now in the final process of being finalized.

Now to conclude, this summary of Irbesartan of adverse event reports is presented in abbreviated format, that there are no concerning safety signals. The use in the pediatrics population is small, and with a few adverse events reports. This completes the adverse events reporting as mandated by BPCA. The FDA recommends routine monitoring of Irbesartan for adverse events in all populations. Does the Advisory Committee concur?

I'd like to acknowledge members of the Office of Drug Safety, and the Office of New Drugs for their help in preparing this presentation and for their guidance. Thank you.

DR. NELSON: Should we go ahead to the Sibutramine?

DR. IYASU: Yes. The next presentation is adverse event reports for Sibutramine. There are two presenters. Dr. Eric Coleman, who is from the Division of Metabolic and Endocrine Drug Product; his background includes residency training in internal medicine, and fellowship training in the areas of metabolism and nutrition. He joined the Division in 1995, and has been a medical team leader since 2000. And the second presenter is Dr. Hari Sachs. Dr. Hari Sachs is a Medical Officer in the Division of Pediatric Drug Development. She is a Clinical Professor at George Washington University, and continues to practice pediatrics. And both these presentations will be on Sibutramine. First, Dr. Eric Coleman.

DR. COLEMAN: Thank you and good morning. I'm going to spend about 10 minutes to give you a very brief overview of the regulatory history of Sibutramine, with a focus on the agency's response to a citizen's petition that was filed back in the spring of 2002.

Sibutramine is trade-named Meridia, and this is a norepinephrine and serotonin reuptake inhibitor that was approved for the treatment of obesity in adults in 1997. In trials of one-year duration, the placebo subtracted weight loss was roughly 3-5 percent, and the principal safety concern with Sibutramine has always been its effects on blood pressure and pulse. The average changes in blood pressure are on the order of 1-3 millimeters, certainly not staggering, but there are subsets of patients who experience much larger increases in blood pressure and pulse.

Because of this concern, when the drug was approved, the labeling did include warnings that the drug could substantially increase blood pressure in some patients. And, therefore, all patients treated needed to have regular monitoring of their blood pressure and pulse. Furthermore, given that the drug does have some pathomimetic properties, it was recommended that patients with a history of coronary artery disease, stroke, arrhythmia, CHF, not receive the drug.

In June of 1999, the agency issued written requests for pediatric studies. One was a standard pharmacokinetics investigation, and the other was a 12-month placebo controlled trial, which I believe you'll hear more about from Dr. Sachs. Briefly, this study involved 450 adolescents with body mass indices of at least two units above the mean for the 95^th percentile. The primary efficacy end-point was a change in BMI. And again, given the concerns about the blood pressure and the pulse, there were subgroups of patients who underwent ambulatory blood pressure monitoring, as well as echo cardiography evaluation.

Skipping to a different topic, in March of 2002, the Italian Ministry of Health temporarily suspended the marketing license for Sibutramine because of the death of a second young woman who was taking the drug for weight loss. Shortly after that in this country, FDA received a citizen's petition requesting that we take the drug off the market due to its unfavorable cardiovascular profile. And a year or so after that, that petition was supplemented with another petition that said there may be concern for fetal toxicity, so we addressed these two issues the best we could, and I'll discuss that.

Before I get to that, the Europeans, at least in this case, were much quicker than we were. In June of 2002, just three months or so after the original event in Italy, the EU regulators got together and re-evaluated the available data, and concluded that the risk benefit profile for Sibutramine was still favorable, and they recommended that the drug be re-instituted back in Italy, which it was. Meanwhile, back here reviewers from metabolic and endocrine were working with reviewers from the Office of Drug Safety, focusing on reports in AERS that had to do with cardiovascular toxicity and fetal toxicity.

Our first focus was to look at the number of deaths in AERS. That certainly is the most objective end-point. And from the time frame, November of `97 through August of 2003, there were a total of 54 domestic deaths reported to AERS. Thirty of these deaths apparently involved a cardiovascular event. I've showed this figure because I think it tells you quite a bit.

This shows the reports to AERS on the Y Axis by calendar year, and quarter of calendar year. And you can see that the reports of death to AERS were relatively high the first two years post-marketing, and then there's a rather dramatic decrease in the number of reports. And then you see this staggering increase in 2002. It turns out this coincides with the negative publicity that surrounded the activity in Italy, and in this country. The numbers gradually decrease down after this, but this is not an unusual finding when there's a lot of press about a drug. There were quite a few lawyer-driven reports in here, I would imagine.

Turning to cardiovascular events, we again focused our attention on serious, non-fatal cardiovascular events, same time period - 1997 through 2003. There were a total of 224 domestic non-fatal CVD events. These events were many and varied. I've listed some here, viral cardiomyopathy, MI, A-Fib, hypotension, cardiac valve disease, just to name five. The important thing to remember is the reporting pattern for these events was very similar to the pattern for deaths. The rates were highest in the first two years after marketing, which is not unexpected. There was a dramatic decrease after that. And then following the negative publicity, we see a major spike in the reports of events.

Now as many of you know, passive drug safety surveillance systems, such as AERS, are most reliable for picking up serious previously unrecognized rare events. Aplastic anemia would be a good example. These systems are not well suited, however, to assessing whether a drug increases the risk for commonly occurring adverse events in the population for which the drug is approved. And this is clearly the case we are in with Sibutramine and cardiovascular disease. Given that the drug does have some patho medic properties, it's not inconceivable that it would increase some people's risk for cardiac disease.

At the same time, obesity itself is a risk factor for cardiovascular disease, so trying to determine causality from a database like this is extremely difficult, if not impossible. Fortunately, there is an ongoing trial called the Sibutramine Cardiovascular Outcome Study, or SCOUT, which is a randomized placebo-controlled trial involving roughly 9,000 obese patients who are at high risk for cardiovascular events. And weight loss and CVD events are two of the primary outcomes of interest. This is an event-driven trial, so it's difficult to say when it would be completed, but it may take up to five years. But if it is successfully completed, it will no doubt provide the rigorous assessment of Sibutramine's safety profile.

Just quickly turning to fetal toxicity, again, the folks in ODS combed through AERS looking for terms that could be related to fetal toxicity. The time span for that search was November of `97 through October of 2004. There were a total of 34 domestic reports of pregnant women exposed to Sibutramine, and of those 34, five were actually reported as specific birth defects, three of which were involving the cardiovascular system.

Now we know from the pre-clinical data that were conducted, the trials conducted, studies conducted before the drug was approved, that there was no signal to suggest that Sibutramine was toxic to the fetus or that it was teratogenic, so we do have animal data going back before the drug's approval that does not raise any suspicion that this would be a problem. We also did a literature search, and found a handful of reports where women had been exposed to Sibutramine during pregnancy, and they reportedly delivered healthy infants. So all together, we felt that Sibutramine's current pregnancy Category C was an appropriate designation that did not need to be changed based on the information that I just discussed, and I show you here the actual language from the Sibutramine labeling.

And finally, although the Agency denied the citizen's petition to take the drug off the market, during the review process we did note, and this is not unexpected, that there were patients who reported cardiovascular events, who apparently had a history of heart disease, or arrhythmia, or stroke, and clearly those are not the people that you would want taking this drug. So Abbott Laboratories, the manufacturer of Sibutramine, did implement some risk management steps to try to enhance more appropriate use of the drug. And these measures, again, were geared towards reinforcing the need to monitor blood pressure and pulse in all patients who take the drug, and also to try to limit its use in patients with known coronary artery disease. I won't go through this list, but again, this lays out the time-line and some of the actual steps that were taken involving labeling changes, Dear Healthcare letters, doctors, and some alterations in the way the medication is supplied, so that people are less likely to get multiple months of drug at one time, and I think I'll leave it at that.

DR. SACHS: Good morning, everybody. Sibutramine, or trade-name Meridia, is an anti-obesity agent that's marketed by Abbott. It was originally approved in 1997, and exclusivity was granted in 2004 for trials in adolescents. The mechanism of action is similar to that of SSRIs. This is a norepinephrine, serotonin, and dopamine reuptake inhibitor, and that becomes real important. Indicated for the treatment of obesity in adults, along with appropriate diet for patients with a body mass index over 30, or for a lower BMI if there is risk factors, such as diabetes, or dyslipidemia, uncontrolled hypertension that are present. The adult dosage, as Eric mentioned, was 5-15 milligrams per day.

Now as Dr. Coleman has shown you, there's been a lot of review of both the efficacy and safety of Sibutramine, starting with an Advisory Committee in 1996, and some reviews by the Office of Drug Safety for cardiovascular and fetal toxicity. This was actually updated again in 2004, and that brings us to today where we'll look at the adverse events once again, specifically in pediatrics.

To put that in a little context, we're going to look quickly at the use. And prescriptions for this class of agents has actually been decreasing over the past three years. Sibutramine use accounts for about 10 percent of the total drug use, of which only 1 percent is pediatric, approximately 4,000 prescriptions. Most of the use in children is in adolescents with the diagnosis of polycystic ovaries or obesity.

We're going to look real quickly at the trials that were performed for exclusivity. And there were two trials, pK and safety study, and an efficacy and safety study. The pK study was a single-dose study that was performed in a sub-population of about 90 adolescents. It found that the pK parameters for Sibutramine and the active metabolites were similar really to that of - between adults and adolescents. The efficacy trial was a placebo-controlled trial of obese adolescents that had appropriately elevated body mass index. The primary end-point for the trial was the absolute change in body mass index.

When the trial was actually reviewed in detail, the zero height measurements revealed that there might be some concerns about the reliability or accuracy of the height measurements, since approximately 12 percent of patients lost height. And for that reason, since the height is so essential for determining the body mass index, and that's the primary end-point, it was felt there was no conclusions that could be drawn from the trial, and the labeling was changed to show that data are inadequate to recommend use in adolescents.

Now when you think of this, please remember when we look at the trials for exclusivity, that is done before there is a really in-depth review of all the data. The trial, as Dr. Coleman mentioned, looked at cardiovascular safety. There was ambulatory blood pressure measurements performed on a subset of adolescents, and despite the fact that the patients did not get medication on the day of treatment, there was a modest increase in blood pressure noted, as well as an increase in treatment emergent hypertension. But as you've heard, there is a bolded warning in the label which states that there can be a substantial increase in blood pressure and heart rate in most patients, and the need to monitor patients that are having therapy. The echocardiography findings did not raise any concerns or detect any valvular anomalies in the over 100 patients that it was performed on.

Now since this agent is related to the SSRIs, there were some psychiatric adverse events noted in the trial, and so labeling was changed to reflect that there may be a potential risk of psychiatric adverse events, and the mechanism of action and the need to monitor patients is detailed. So as a result of the studies performed for exclusivity for pediatric use, section has been changed. This is the labeling in full, and it briefly says that the efficacy in adolescents who are obese has not been adequately studied. There's data about the trial, the psychiatric adverse events, the need to monitor the patients, and then the labeling concludes that it is inadequate to recommend use.

We'll look briefly at the safety labeling, which is pertinent for the pediatric patients and the adverse events that were seen since approval. As I mentioned, there's a bolded warning about the potential risk of increased blood pressure and heart rate. There's also a warning to avoid use in patients with cardiovascular risk factors, and to exclude secondary causes of obesity.

Cautions also describes that seizures might occur, and to use cautiously in patients with bleeding diathesis. And as Dr. Coleman mentioned, this is a pregnancy Category C drug, which is based on the fact that there may be adverse effects on the fetus in animal studies, but there's limited data on this.

Since market approval in 1997, there have been almost 6,000 total adverse events, of which 1,000 of these are serious, and 118 were deaths, although only five of these deaths were in pediatric patients. Remember, these numbers include duplicates, and when we look at the data, we often find there's a lot of duplicated reports.

Now the five deaths in children were related, two of them were cardiac, and three of them were related to pre-term. And as was mentioned, there was a very detailed analysis of all of the adverse events, including deaths, which didn't show a pattern. There were 24 unduplicated serious non-fatal adverse events since market approval. Nine of these were associated with overdoses. There was one case each of Qtc prolongation, which occurred during the exclusivity period, and I'll talk about that momentarily. One case of a seizure, non-insulin-dependent diabetes, granulomatous uyeitis. And again, there were some congenital anomalies that were noted. As we've mentioned, the underlying events you can see may not be specifically labeled, although some of them are certainly related to labeled events.

Now during the pediatric exclusivity period, there were approximately 150 total reports, of which 140 were serious, and there were 18 deaths in all ages, but no pediatric deaths. The one pediatric adverse event was a 14-year old patient who, during clinical trial, had Qtc prolongation. He had a baseline abnormality on his EKG, but the Qtc did rise above 450.

So in conclusion, the labeling for this drug has been updated after the exclusivity studies. The one-year review did not really reveal any new pediatric adverse events, or repeat pediatric adverse events. There's been multiple reviews by the Office of Drug Safety and the Division, which doesn't really show any increased cardiovascular or fetal risk. You've heard about the SCOUT study, which is ongoing, to formally evaluate the non-fatal and fatal cardiac events in patients. And we propose monitoring this drug for one-year, just because there was only one event during exclusivity period, and a small number, if you all concur.

Once again, I'd like to thank all the folks who participate in these reviews. There's a lot of folks behind the scenes.

DR. NELSON: Thank you. Do we have comments from sponsors? Okay. Hearing none, we have three questions, and why don't we just go through them one-by-one and address any particular questions that people may have. And the first is for Clofarabine; the recommendation is routine monitoring. So is there any discussion or questions? Judith.

DR. O'FALLON: I'm not sure whether I'm reading this correctly, but it looked to me like there's been only seven and a half months of observation since the thing went out on the market, and that the Children's Oncology Group hasn't reported any yet who are on the study, so it seems to me that we have a small sample. This is the one where the exclusivity occurred before the thing went out on the market by about five months, and I just wonder if that gives us a big enough sample. That's the only issue. It had a number of deaths, and so on.

DR. GRYLACK: You're correct, Dr. O'Fallon, about the unusual juxtaposition of the marketing approval and exclusivity. As to whether that's a sufficient period of time, I can't say for sure. I don't know whether anybody from the Office of Drug Safety or Oncology want to comment on that, but certainly that would be a concern, especially since the usage, and the usage is just an estimate by the sponsor at this point, is limited due to the nature of the indication.

DR. O'FALLON: This is a small sample issue, and if you're really looking for signals of rarer adverse events, the smaller the sample, the less chance you have of getting them. That's what my concern is.

DR. NELSON: I guess it would be useful to know if COG has any plans to incorporate this in trials, because if they don't currently, then the usage is going to stay very low, and we may not know that at this point.

DR. MURPHY: We did talk to them. Solomon, do you remember if they mentioned any -- because this very low use, we did call them to see if they had any additional data.

DR. IYASU: I think the information we got from them was currently there are no open studies, so they were not able to provide us information on use of this medication in pediatric patients. So the only information we have about use is from the sponsors, which is an estimate based on the sales data that they have.

With respect to the question about the small sample size in AERS, when we say we're going to do additional monitoring or returning it to routine monitoring, it does mean that the Office of Drug Safety is going to monitor, just like for any adverse events for a long time, as far as the reports that are coming to the FDA. The difference with other drugs that we've looked for an additional year is that we do even intense or in-depth look at the adverse events for an additional year, because we have a concern that is not being adequately answered by the numbers that we see in the AERS system, so it doesn't mean that it falls off the radar screen when we say we'll look at it.

DR. NELSON: Thank you, Solomon. So let me just ask if there's anyone who would disagree with routine monitoring for this particular drug at this point, which routine is actually as intense as it always is. In the interest of time, I'm not going to take a roll call. I'll just note for the record that the only individuals, apart from the members of the Pediatric Advisory Committee, who are voting in this section are Paula and Bob, since the consultants for the ADHD portion have not been screened for conflict of interest surrounding these issues, for the record. So I'll note that that is unanimous in agreement there.

The next is the Irbesartan. Any comments or discussion about that issue? I'll just make one note about the label. I know we've had discussions on this in the past about what information goes in when there is no pediatric indication. I would ask that at least the FDA consider the possibility that the information on the top slide on page 18 be in the label, which is quite helpful to tell the pediatrician that, in fact, it's not been shown to be useful. I know that at times putting negative data in labels is felt to sometimes suggest that people ought to use it, but I think that -- I doubt this information would be found anywhere else if it's not in the label.

DR. SHAPIRO: I'd like to comment on that. The pharmacokinetic data is being pulled from the label, but in the proposed labeling it's going to say Irbesartan in a study at a dose of up to 4.5 milligrams per day once daily did not have a lower blood pressure effective in pediatric patients ages 6-16 years of age. So essentially, there is going to be information in the label describing that it was studied and that it was not effective. We're just pulling out the pharmacokinetic data because we don't feel -- it doesn't get the indication.

DR. NELSON: Well, thank you. That clarifies it. Bob.

DR. TEMPLE: That reflects a general commitment to put the negative results as they occur into the pediatric sections or the indication section. One doesn't want to overstate the negative. Sometimes absence of a positive doesn't really prove something, but it will say as truthfully as possible what's been found or not found.

DR. NEWMAN: Just to say that I didn't hear a mention of the sample size of the study going on the label, so probably the best thing to do would be to give the point estimate in the 95 confidence interval for the main effect on blood pressure, so the person reading the label can tell how much actual data there was addressing that question.

DR. NELSON: I seem to recall you make that point every meeting we have, Tom. Any other comments on Irbesartan? So I'll ask again the question, the negative, anyone who would disagree with returning this to routine monitoring? So seeing none, I'll note that the vote is unanimous for that recommendation. And now on to Sibutramine. Comments, questions from the Committee? Tom?

DR. NEWMAN: I notice that once again we have medication that was granted exclusivity based on a study that FDA on further review decided was inadequate. And Dr. Sachs mentioned that the exclusivity decision was made before they reviewed the data in depth, but just I think this is troubling, and maybe they need to review the data in depth before they grant exclusivity, or revoke the exclusivity or something, but it does seem troubling to give exclusivity and then have the FDA say that the trial wasn't adequately done.

DR. MURPHY: As you know, we recognize this issue, and we think that it is problematic. And we can't revoke the exclusivity, and actually it's sort of, by the time you do it they've gotten most of what they need, so the issue is developing an approach which - in the re-authorization, maybe, Congress will reconsider that - where they could still get exclusivity even if there's a negative finding, but that we have more time before we have to make that assessment. And I think the Europeans are moving in a direction in which they are going to have more time to do that, so I think it might be something for us to consider, also.

DR. NELSON: Other comments? John.

DR. MOORE: One concern I have is related to presumably this one slide that really talks about the results of adverse events in fetus. Two of the adverse events are hypoplastic left heart syndrome patients, and just a quick calculation, I would expect one HLHS patient to arise in about 10,000 unselected pregnancies, so I'm wondering if this is something that is a concerning signal. I don't know what the estimated use in pregnant women was. I know it's warned not to use it in pregnancy, but I suspect that there may be an issue there, there could be.

DR. NELSON: Bob.

DR. WARD: John, I saw the same thing, but there are so many issues that I think also have to be addressed that may not be recognized about familial association of hypoplastic left heart, or Turner Syndrome, or things like that, that I think the current approach of continued intensive monitoring is more justified than taking an action on limited information presented here. I also wanted a clarification. I assume that myoplastic left heart syndrome is what we would refer to as hypoplastic left heart syndrome. Is that correct?

DR. SACHS: This was a foreign report, and that's what it says, so I can't confirm that or deny it - don't know.

DR. NELSON: Well, for the purpose of our discussion, I would point out that the recommendation is, in fact, the monitoring continue at the higher level of scrutiny, so hopefully over time some of this will become clearer. In the interest of time, I'd just ask if there is anyone who would disagree with continuing more intensive monitoring for this particular drug, which I assume would then come back to us at some point in the future with a report.

DR. MURPHY: I think if that is what you would wish, it might be helpful to say, based on X number of more cases or a period of time, just give us some idea, because you've done that in the past, of when you would like us to come back, if that's where you're going on this.

DR. NELSON: I mean, I guess I'm comfortable leaving that open to some judgment. I mean, it depends when the SCOUT study is done and when data comes in. I mean, some of it depends on the accumulation of data, which would be hard for me to say in the abstract, 12 months, 18 months, 24 months.

DR. MURPHY: That's the sort of thing I wanted you to say. It's not a year or two years.

DR. NELSON: I'm sure Tom would want a confidence interval with perhaps some of --

DR. MURPHY: You want us to come back after we have additional data, particularly from the SCOUT and additional reporting that we think might help us assess this.

DR. WARD: Could we ask that any cardiovascular events be described in as much detail as possible, and maybe be investigated in more depth, especially as part of this SCOUT study.

DR. NELSON: Other comments? So again I'll ask you in the negative, anyone have any objection to more intensive monitoring over the next year or two, depending upon as data comes in? So I'll note that the vote is unanimous on that, as well. And that brings us now to ADHD. Diane.

DR. MURPHY: My task is to first take you back to June of last year, and then to bring you back to today, to help you transition to the focus on the treatments for ADHD.

We're going to be asking you to talk about risk, so just to remind everybody that there is - this is double negative, Dr. Temple. I understand that, but that there is no drug without some risk, but that an untreated condition also carries risk. A successful treatment depends upon many things, but these are some of the things that we will be focusing on because the FDA and the panel cannot impact all of these, but we certainly have other academic groups or organizations that might help us make sure that we have proper diagnosis. The job of making sure we have the right dose and the therapy lie with the sponsor and FDA, and what we're going to focus on today is our knowledge about adverse events adequately expressed; are there better ways of expressing it, or do we need to do anything about it. In other words, an appropriate communication about risk is what we are asking the Committee to address today.

So back to June 2005 - some of you weren't here, but many of you were. And at that time, the FDA brought forth for its review under Section 15, Concerta, and we expanded it to the other methylphenidates at that time. We said we had identified two possible safety concerns with the methylphenidate products, which were the psychiatric and cardiovascular events, and that we intend to make labeling changes, but that we wanted to examine all, and at this time we said stimulants, but we meant any of these products that are being used to treat ADHD. And that we asked you to basically advise us if you were in agreement that we do that, because we did not want to direct people improperly from one set of products to another until we had looked at all of these products. So then we were examining the post-marketing reports, and we would come back to you in early 2006, and I think we qualify March as early 2006. And that at that time, for the cardiovascular events, we said the Agency believes it's not yet possible to determine whether they are causally associated. We are pursuing additional means to better characterize the cardiovascular risks for all the products approved for ADHD, and we propose that we obtain additional data to help guide us, and we're going to update you today on where we are with this issue.

Now in your background package, instead of sending you everything from June again, what we tried to do for you is summarize the psychiatric and neurologic events for the methylphenidate products that were in your package, because those were the ones which we had the most extensive summary for you, so that you'll find one table in there that is a repeat from June.

Today's update on pediatric safety, this is a summary of all the data. It's not in sequence with the speaker, so what you're going to get is the review of the AERS data for Adderall XR during the one-year post-exclusivity. You're going to get -- you received and will have presented to you in a condensed form the review of the AERS data for psychiatric events for most products used to treat ADHD, including additional reports requested from the sponsor.

What we're trying to tell you is we tried to get as much information as we could. We went back to the sponsors, the Division of Psychiatry and Office of Drug Safety, sent letters to the sponsors, asked for additional information. That information was submitted, and it has been analyzed, and so that is in your packet, and will be reviewed. And in that process, there is a review of the sponsor's submission of psychiatric adverse events from 100 ADHD clinical trials in response to this request, so again, we've tried to maximize the amount of information we can provide you, and a review of the cardiac adverse events from AERS for the ADHD products. And we will be talking about possible future studies to assess cardiac risk. So that is our update that you will be hearing today.

And your job is to assess the potential psychiatric cardiovascular risks with the use of products to treat ADHD in the pediatric population, with the best information that we can provide you at this time; to advise FDA on how best to communicate risks so the physicians and parents can make informed decisions. This is really the crux of where we're trying to go. And advise FDA on how best to address risk for all the therapies used to treat ADHD to avoid inappropriate switching to products for which we have less safety information, so we'll be asking you to look at all of these products.

Now the questions - I am not going to read these in detail, but basically the preamble, if you will, that because these products have been shown to be effective when used to treat children who have been properly diagnosed as having ADHD, there will continue to be a need for access to these therapies. You have received information on potential cardiovascular and neuropsychiatric risks associated with the use of medications to treat ADHD in children. This information included adverse events from spontaneous reports and clinical trial data, the current labeling, and FDA's plans for additional studies on cardiovascular risk.

In consideration of this information, we're asking you to address the following questions concerning psychiatric and cardiovascular adverse events. And we have broken up the questions into two sections; the psychiatric adverse event sections - and I'm not going to read them because they will be read to you by the Chair as you go through them - and the cardiovascular. It's the same set of questions addressing those adverse events.

And I want to thank all of you very much, and I particularly want to comment that I know not only is there a lot of information to read, but a number of you have had to cancel other commitments and to make time in your busy schedules, and we sincerely appreciate your efforts to be here today and to advise us. Thank you.

DR. NELSON: Perfect. Thank you, Diane. Our next presentation is going to be on labeling. And I might point out to the Committee that this might at times seem to be a dry topic but, in fact, the label is the primary mode of communication, and the questions we've been asked to address is precisely communication, so take notes because we'll be talking about the labeling at the end of the day.

DR. MURPHY: And I was supposed to introduce Dr. Rosemary Roberts, who is a pediatrician and Board Certified in Pediatric Infectious Disease, has been at the Agency I won't tell how many years, Rosemary, but is basically one of the founding mothers, if you will, of the pediatric program at FDA, has an extensive experience in not only pediatrics, but in the regulatory milieu, which we need to develop these products.

DR. ROBERTS: Thank you very much, Diane, for the nice introduction. Good morning, everyone. And yes, they gave me the dry topic, so now, just when you thought you knew how to attempt to read one of our labels that are pages long, I'm going to tell you that the labeling in its format and its order are being completely changed; so, hence, where will the safety information be in the new format of the physician's labeling?

Okay. Well, first of all, just some simple definitions. The label is the written printed material, graphic material on the immediate container of the product, whereas labeling includes not only the label, but all the written and graphic material that goes with the product. So if we look here at Strattera, the label is specifically what is on the immediate container, and then what you see in the back is the fine print physician's package insert. Who knows what labeling will look like as we move into the electronic age?

Now labeling is to be a summary of essential scientific information needed for the safe and effective use of the drug. It is to be informative, accurate, and not to be promotional in tone, nor false or misleading, and based on data derived from humans, wherever possible.

Now a little bit of background as to how we came to the new physician's labeling. Previous labeling regulations date back to 1979, and over the two decades from `79 into the early 21^st century, there was increasing amount in complexity of drug information that was being put into the labeling. So the goal was to have more informative and accessible labeling, resulting in better risk communication and management. And in order to do this, the process included focus groups, a National Physician Survey, public meeting, and written comments. And armed with all this information, a group of people at the Agency worked on the proposed rule for Physician's Labeling that was issued in December of 2000 and then the final rule, which was published in January of this year.

Now there are certain innovations that I want to make you aware of. There's a highlights section, where it's just going to the very short bulleted statements about key sections of the new labeling. There's a table of contents, what an original idea when you have something that's that long. But not only is there a Table of Contents, but you will be actually hyperlink to the direction section that you want to read.

There will be a section that identifies the recent major changes in the labeling, and it will also indicate the original date of approval within the United States. So this is an example of a fictitious drug, and I'm going to use this fictitious drug to highlight the safety areas as I go through this. But as you can see, there is -- at the very top it says that this is limited information, and you will have to go to the full prescribing information as you move on.

Now this is the old format of the labeling. You're probably familiar with this. The first thing you see is the chemical structure, which for most of us doesn't do much for us, and we usually don't spend a whole lot of time there, nothing against chemists. Okay. So on this highlight, these are the things that you're going to see when you first open up either a hard copy or the electronic copy of the labeling. So at the very top it's going to say that this is limited information.

I'm going to introduce the drug Imdicon or Cholinasol. Now this particular drug is not going to meet what labeling should. It's not going to be a summary of essential scientific information needed for the safe and effective use of the drug, because there's no data at all behind it. However, it is going to be useful for demonstrating what this new labeling is like.

Imdicon is an imdinicine diphosphate antagonist platelet aggregation inhibitor. It's indicated for reducing the risk of thrombotic stroke whenever there's been precursors of thrombotic stroke within the individual, or they have had a completed thrombotic stroke. And it's also indicated for reducing the incidence of sub-acute coronary stent thrombosis.

You can see for this particular drug it was approved in 2000. There is a boxed warning. You will see the recent major changes there on the left-hand side of the screen under the box, the indication in usage and the dosage in administration. As we move to the right-hand side, the dosage forms and strengths, bulleted contraindications, warnings and precautions, adverse reactions, where to report suspected adverse reactions as we're trying to encourage more people to report through the MedWatch system, the drug interactions, and use in a specific population.

You will also note, for instance, if you look at the warning which, is probably the easiest to read from a distance, under netropenia agranular cytosis, the first bullet, it says 5.1. That tells you that if you want to read the full prescribing information on that, you go to Section 5.1 of the labeling. Okay.

Now this is the Table of Contents. This is going to be the order in which you see the information in the labeling. So as you can see, the boxed warning is at the top, you immediately come to indications and usage, the dose that you want, the dosage forms and strengths, and then you start into the safety information. Once you know what the drug can be used for, how it's to be administered, then the contraindications, warnings and precautions, and adverse reactions. So here's the Table of Contents you will be able to take, and if you have this in electronic format, click on warnings and precautions and immediately be hyperlinked to that section, so much easier than scrolling down or trying to find the right page or reading one of these that's very long.

Okay. For contraindications, for all of you who concentrate on numbers, I want you to know that the Section 201.57, which is the labeling section of our Code of Federal Regulations, now has been reordered, so that, indeed, contraindications is falling in this Subsection C-5. Contraindications are to describe those situations in which the drug should not be used because the risk of use clearly outweighs any potential benefit. These are to be known hazards and not theoretical possibilities.

Now the other thing, there is a very good guidance that was actually a draft guidance that was put out and is available to you that talks about each of these sections. And its contraindications are divided up into observed as well as expected reactions. And these would be expected reactions based upon what we know about the drug class or the drug pharmacology, or there are certain clinical situations where you would highly suspect that a product such as this one, which is used to prevent clotting, if you had a patient with severe hepatic impairment, they are more likely to have bleeding problems and, therefore, this drug is, indeed, contraindicated in people with severe hepatic impairment.

And from our fictitious drug, Imdicon, the contraindications are not surprisingly in those that have hemtopeioitic disorders or a history of thrombotic thrombocytopenic purpura or aplastic anemia, those with hemostatic disorders or active bleeding, and those with severe hepatic impairment.

Moving on to warnings and precautions - again, this is divided into observed and expected, and it would include clinically significant adverse reactions observed in association with the use of a drug for which there is reasonable evidence of a causal association between the drug and the adverse reaction. The causal relationship need not have been established, and I've underlined observed because this means that they have been observed, and I've underlined causal because that's a new word that's been introduced into the Code of Federal Regulations from our previous.

Now, so what kind of observed reactions would fall into warnings? Well, if it's a serious adverse reaction. Now serious, fortunately they didn't change the definition of serious. It's still where the outcomes would be death, some kind of life-threatening adverse event where a patient had to be hospitalized, or if they were in the hospital, they had to extend their hospitalization, congenital anomalies, et cetera.

What are otherwise clinically significant adverse reactions? These are adverse reactions where it would require discontinuation of the drug, where one would have to make a dose regimen adjustment in order to use the drug safely or there would be a significant effect upon patient compliance. For instance, if you had a drug that had a photosensitivity reaction and you're going to prescribe it for a person who works outside all the time, well, you may want to reconsider that. That person may stop taking that drug because of their job, so try to look for another drug, or a product that interferes with a laboratory test.

Now, then there are adverse reactions which can be expected to occur with a drug based upon observations from other members of the drug class, or the animal studies. These would be serious, or otherwise clinically significant adverse reactions that either would be based upon what is known about the pharmacology, chemistry, or the drug class, and therefore it's likely that an adverse reaction could occur, or animal data raises substantial concern for the adverse reaction in humans.

Other factors that need to be considered for placement of adverse reactions into this section would be the indication. For instance, if you have a self-limited condition where you have adverse reactions such as alopecia or severe nausea, you may want to consider putting that in this section because that would potentially lead to a person considering discontinuing this drug because they have such a self-limited condition; whereas, if it was a chemotherapeutic agent, those kind of issues may not warrant putting it into this section of the labeling.

Also, the incidence, if there's a high incidence of the particular adverse reaction or if there -- you could actually manage or prevent the adverse reaction through some kind of specialized patient monitoring or titration of the drug, et cetera. So here's an example of the adverse reactions that went into the warning section for the fictitious drug, Imdicon. And as you see, there's neutropenia at a 2.4 percent incidence. It also tells you it may occur suddenly. It will typically resolve within one to two weeks of discontinuation. This is very important information to know that if you discontinue the drug, you can expect it to resolve. And then the second bullet talks about how to monitor for the particular hematologic adverse reactions. Again, you can see that you're then referred to Section 5.2 of the full prescribing information.

Now the boxed warning is simply a subset of warnings, and it's where there is information that we want to highlight for the prescriber, and therefore it is put in a very prominent place within the labeling and within a box. And when should one use a boxed warning, or when do we contemplate using a boxed warning? If the adverse reaction is so serious in proportion to the potential benefit from the drug that it is essential that the adverse reaction be considered when one is assessing the risks and the benefits of using the drug. So again, it goes back to the benefit risk equation, and where does that lie?

The serious adverse reaction that can be prevented or reduced by appropriate use of the drug to be put in this section or drugs that are approved with restrictions to assure that there is safe use in either distribution or use restriction.

Ordinarily, information that goes into the boxed warning is based upon human clinical data. However, serious animal toxicity can be enough to introduce it into a boxed warning. So again, this just shows you the highlighted boxed warning. This is not the full boxed warning. This is only bulleted highlights from that section. And that, again, from our example from the Imdicon, it has information on how to monitor for the hematologic adverse reactions, tells you when to discontinue Imdicon, and the adverse reactions for which you would want to discontinue the drug.

Now, how do we go about knowing when considering a boxed warning? And it really is a balancing act. It's balancing the risks of the drug versus the benefit of the drug. And I want you to concentrate on this and think about it because it's a very serious situation that we consider with a great deal of knowing that the consequences of what the black box or the boxed warning can lead to, consequences for patients, consequences for physicians, consequences for sponsors, so we don't take it lightly.

It is a multi-layered process, and by that I mean internally we involve our statisticians, toxicologists, medical reviewers, safety reviewers, clinical pharmacologists, every discipline that's appropriate within the Agency. We often bring it to groups such as yourself, external advisory committees, for input. It's ultimately a consensus decision. It's not a unanimous decision usually, but it's a consensus decision, and it's a very iterative process in which the sponsor plays a role.

And here are resources that are on the FDA's website, it talks to you about the new labeling. There are several fictitious drugs there to give you a handle on how some of the new labeling will look, and hopefully to help us and to help sponsors as we try to put labels in the future into this new labeling.

I want to thank you very much for your attention.

DR. NELSON: Thank you, Rosemary. The next presentation is by Paul Andreason, who is Acting Deputy Director of the Division of Psychiatry Products at CDER.

DR. ANDREASON: Thank you very much. This morning I'd like to bring you up-to-date on what's been going on in the Division in products that are used to treat Attention Deficit Hyperactive Disorder. Just as an update, the June 30, 2005 meeting we discussed Concerta in the BPCA review, and in the discussion we came to the conclusion that even though many of these adverse events were known to the psychiatric community, they perhaps were not well communicated to other prescribing communities, such as the pediatric and general practice groups. And that we would look at psychiatric adverse events in more detail, with the hope of talking about this today, which we will.

Just to outline some of the things that have happened since June of 2005, and I was not completely aware of just how busy we've been, Canada returned Adderall XR to the market on August 24^th, 2005 after removing it in September of 2005. Then the randomized clinical trials for the psychiatric adverse events for all treatments, a review of those was designed and letters were sent to the sponsors in September of 2005 with a deadline of November 1^st for getting that data in to us.

Then in October, Pemoline was removed from the market. In December of 2005, we held a Psychiatric Drug Advisory Committee for the review of methylphenidate transdermal patch, and they recommended approving it, and we completed our review of that and issued an approvable action on the 23^rd of December, 2005.

In January we requested a labeling update that warned about use in patients with known cardiovascular defects or stimulants. This was language that was already in the Adderall XR labeling, and this was the same type of labeling that had been in there prior to Canada removing it from their market. And when they placed it back on the market, it contained that labeling that we already had. We sent that updated labeling request to the makers of all stimulants. And that's already actually incorporated into the Concerta labeling, if you look at their website now.

In March, actually March 3^rd of 2006, the results for the review of the randomized clinical trials for the psychiatric events was filed in our Division file system. Thank you to the Office of Drug Safety people for doing that. And today, here we meet for the Adderall XR BPCA review and presentation of those psychiatric adverse events that we planned to study. And I would like to say that that was a large task, and I'm really proud of the folks that did that, because it was no small feat. And then tomorrow we meet on Modafinil, and its potential use in the treatment of Attention Deficit Disorder with the Psychiatric Drugs Advisory Committee.

For those of you who have not seen the labeling for cardiovascular risk and structural cardiac defects, this is what it says, sudden death and pre-existing structural cardiac abnormalities is the section title, or the subsection title. Sudden death has been reported in association with CNS stimulant treatment at usual doses in children with structural cardiac abnormalities. Although some structural cardiac abnormality alone may carry an increased risk of sudden death, stimulant products generally should not be used in children, adolescents, or adults with known structural cardiac abnormalities.

Now if you go and look at all of the existing labeling for the stimulants that are currently on the market, you will notice that there are some differences where risks are placed in labeling, or whether they're mentioned, or what kind of data is there. That has to do with several factors; first of all, the age of the product, and what kind of data was available or required for the approval of that product at the time, and the breadth of the program. The products for which we have the most data are those that produced information for us as the result of a written request.

In the package for the Committee today, we have a line-by-line, column-by-column comparison of the cardiovascular risk language, but generally speaking, what the stimulants say is that these drugs should be used with caution in patients with hypertension and tachycardia, cardiovascular or cerbrovascular disease, or cardiac arrhythmia, that blood pressure should be monitored, as well as pulse, and that modest increases in pulse and blood pressure have been noted with their use.

For the psychiatric adverse events, I kind of divided these out for the purpose of illustrating the differences of the drugs over time. In yellow, you'll see examples of labeling for the more recent products, and in, I suppose that's purple, you'll see the oldest drug -- excuse me -- for yellow the newest, and purple the oldest. And the adverse events that are focused on in labeling generally revolve around psychosis, agitation, and anxiety.

Now for the amphetamine-containing products, Adderall XR is an example of one of the new ones, Dexedrine is an example of the oldest one -- this is what you see. In Adderall XR you have a warning section that talks about psychosis, clinical experience adjusted in psychotic patients, administration of amphetamine exacerbates symptoms of behavior disturbance and thought disorder. And then the purple section is not only in Adderall, but it's actually in the Dexedrine labeling, too, word-for-word. And then you go on to see some psychiatric symptoms listed in the table of commonly occurring adverse events.

Now some of the concern has been what does the labeling say about toxicity and usual doses. I think most people are aware of the psychiatric cardiovascular events that occur in overdose, but some of the concern that has been raised not only in the June 2005 meeting, but since then, is about these risks that may occur, or what are the potential risks of these occurring at usual doses.

In the Dexedrine labeling, under the overdose stage, in the overdose section it says, "While toxic symptoms occasionally occur as an idiosyncracy at doses as low as 2 milligrams, they are rare with doses of less than 15 milligrams", and then it goes on to say that, "30 milligrams can produce severe reactions." Now Dexedrine is the amphetamine, the active ingredient in the mixed salts of amphetamine, so roughly if you double the dose here, that would be the Adderall XR dose. So that gives you an idea of what the therapeutic index is here.

The Adderall XR labeling says "toxic symptoms may occur idiosyncratically at low doses." So the information is there, but not necessarily as accessible as maybe one might have it. Again, here's the Adderall XR overdose section, starts out by saying "individual patient response to amphetamine varies widely, toxic symptoms may occur idiosyncratically at low doses." However, this is in the overdose section.

The Dexedrine labeling, even though it's the oldest, is amazingly data-rich. "Individual patient response to amphetamines varies widely", and this is the same phrase that I just read to you, but it goes on to say, "30 milligrams can produce severe reactions, yet doses of four to five hundred milligrams are not necessarily fatal." And then it goes on to talk about some animal data, and what the manifestations of overdose may be.

For the methylphenidate-containing products, this is kind of the same breakdown. I'm comparing here Concerta versus Ritalin. Again, there are sections on need for comprehensive treatment program, stating that the drug alone is not an adequate treatment for Attention Deficit Hyperactive Disorder. Again, there's a section in the Concerta labeling about warnings of psychosis, and toxic psychosis has been reported.

Now this was a piece of confusion before that many people felt that toxic meant overdose; whereas, toxic psychosis is a term of art in psychiatry, meaning that it's not functional psychosis necessarily caused by a psychotic disorder, but by a substance or something introduced into the body.

And then in the patient package insert for Concerta, it talks about risks of increased blood pressure and psychosis, and talk to your doctor before taking Concerta if you have -- and then it lists the symptoms. These are some more comparisons between Concerta and Ritalin. Concerta is contraindicated in patients with marked anxiety, tension and agitation, since the drug may aggravate these symptoms. And again, in the patient package insert this is outlined.

There's one statement in the Ritalin labeling that is not in the Concerta labeling. However, I don't think it necessarily adds that much. "Patients with an element of agitation may react adversely, discontinue therapy, if necessary." And then we move on to atomoxetine, a non-stimulant, and this is the information that's in that labeling. There is actually a boxed warning for suicidal ideation. And there is a Med Guide that is also part of the atomoxetine labeling.

The warning sections, it talks about the symptoms that have been reported, including anxiety, agitation, panic attacks, insomnia, irritability, hostility, aggressiveness, impulsivity, akathisia, or psychomotor restlessness, hypomania, and mania. And then a statement on aggression, call your doctor if treatment emergent, and it talks about the incidence that it occurred at.

Now one of the things that came out of the February 5 meeting, or as we discussed the February 9 Drug Safety and Risk Management meeting, was that some people felt that perhaps there was some confusion about whether or not atomoxetine was being considered by the Committee as a stimulant or not, and that there may have been some confusion about nomenclature. The amphetamines and methylphenidate products are accepted as central nervous system stimulants. Atomoxetine is marketed as a non-stimulant.

Modafinil is listed as a wakefulness promoting agent. It has some, but not all, the properties of stimulants, and I just wanted to quickly go through some of those comparisons. Amphetamine blocks, reuptake and norepinephrine and dopamine into the presynaptic neuron and increases release of norepinephrine and dopamine. That dopamine, they say, comes from the reserpine-insensitive cytoplasmic pool, as opposed to the reserpine-non-sensitive granular pool, and in animal studies it increases locomotor activity in mice and rats. Amphetamines have been extensively abused, tolerance, extreme dependence, psychological dependence, and severe social disability have occurred with them, and there are reports of subjects who have increased the dose to many times the recommended therapeutic doses in their abuse. Abrupt cessation following prolonged high dose results in a withdrawal syndrome of extreme fatigue, mental depression, and it is a Schedule II non-narcotic.

Methylphenidate also blocks reuptake of norepinephrine and dopamine in the presynaptic neuron, and increases release of norepinephrine and dopamine. Again, it releases dopamine from the granular reserpine sensitive vesicular pool, and it, too, increases locomotor activity in mice and rats. Chronic abuse has been documented, and it can lead to marked tolerance and dependence, and careful supervision is required during withdrawal from abusive uses. Severe depression may occur. It also is a Schedule II non-narcotic.

Modafinil in vitro binds to the dopamine reuptake site and causes increased extracellular dopamine, but there is no increase in dopamine release. There is increased locomotor activity in animal studies, but it has no effect on the norepinephrine system. Modafinil's reinforcing is evidenced by its self-administration in monkeys previously trained to self-administer cocaine. In humans, Modafinil produces psychoactive and euphoric effects, and it's a Schedule IV controlled substance.

Finally, atomoxetine produces selective inhibition of the presynaptic norepinephrine transporter, or SNRI, blocks reuptake of norepinephrine. It does not stimulate release of dopamine or norepinephrine, and it does not increase extracellular dopamine. It has minimal affinity for either the serotonin dopamine transporters, or other neurotransmitter systems. It does not increase locomotor activity in mice and rats, and drug discrimination studies in rats and monkeys were inconsistent at showing stimulus generation between atomoxetine and cocaine, and there's no evidence of symptom rebound or adverse event suggesting a drug discontinuation or withdrawal syndrome, and it is not a scheduled substance. So that is kind of an overview on the differences between the drugs that are accepted as stimulants and others that kind of have not necessarily been grouped into that category.

Thank you very much. I'd like to introduce Dr. Marsha Rappley from Michigan State University College of Medicine.

DR. RAPPLEY: On February 9^th, the Drug Safety and Risk Advisory Committee met to produce recommendations to the FDA about how best to study the serious rare cardiovascular and cerebrovascular events associated with medications used to treat Attention Deficit Hyperactivity Disorder. Three of us were present, three of the current Pediatric Advisory Committee were present and participated in that meeting; myself, Dr. John Moore, and Deborah Dokken. I'm going to provide a very brief summary of our view on how that meeting transpired.

The FDA presented background on the medications, and the discussion quickly moved to strengthening the warning associated with the use of these medications. The three of us expressed concern about how quickly that discussion shifted to the warning that would be appropriate for the use of these medications, and that it occurred at the expense of an in-depth discussion of the actual risk in children and adults, how that risk differs, and how best to convey that risk to families and physicians.

A vote was called, a yes/no vote to either support or not support the black box warning for these medications in the use of ADHD. And among the three of us, our vote was split. There was one vote in favor of that black box warning, and two votes not in favor of that black box warning. All three of us felt that that vote was called without discussion of content of that warning, or what the wording might be in such a warning, and that there was brief discussion that this black box warning is the highest level of warning that can be associated with a medication approved for use by the FDA.

Further comments from the three of us include from Dr. Moore, that the risk and the benefit of these medications and treatment of ADHD, particularly for adults, has not been well-defined. From Ms. Dokken, that today's further discussion about the best way to convey the level of risk to families and physicians is extremely important. And from myself, that what happened at the previous meeting, I believe is a legitimate concern about the increased use of these medications being confused with the actual risk of the medications themself.

Now I'd like to invite both Ms. Dokken and Dr. Moore to add additional comments, at your pleasure. Thank you.

DR. NELSON: Thank you, Marsha. We're ahead of schedule.

MS. MURPHY: We note it.

DR. NELSON: Well, what I'd like to recommend, because I'm sure we've got a lot more things to do, is we might as well take our 15 minute break now, but what that will mean is instead of starting again at 9:50, I've got 9:22, so how about if we start at 9:35. We'll just shorten it a bit, so be back here at 9:35 instead of 9:50.

(Whereupon, the proceedings went off the record at 9:22 a.m. and went back on the record at 9:37 a.m.)

DR. NELSON: So we could begin to start taking our seats. While Dr. Vitiello gets himself prepared, I have two announcements. So both announcements are actually directed to our esteemed audience. For those individuals who had pre-registered prior to the closing deadline for making comments during the open public hearing, we do need you to still let the folks at the front desk know that you're here. So if you didn't actually sign in, make sure you do that so we know that, in fact, you're here. And the second is, there are some people that are standing in the back. They're certainly welcome to keep doing that if that's their preference, but if you've got a seat next to you that's empty and filled with your briefcase or whatever, you could make room if they'd like to sit down. There does look to be plenty of extra seats. Those in the back are welcome to keep standing, as well, depending upon your back.

So our next speaker is Dr. Ben Vitiello, who's with the National Institute of Mental Health.

DR. VITIELLO: Thank you. My task here is briefly to summarize the efficacy of stimulants in atomoxetine in the treatment of Attention Deficit Disorder before we embark, actually, in talking about the safety issues.

The focus will be on children, and adults for that matter, age six and above. Even though amphetamines are actually marketed and indicated also for children age 3-6, the data in pre-schoolers are much weaker in some respects, so my presentation refers to subjects age six and above.

A word about Attention Deficit Disorder. It's a condition that is characterized by developmentally abnormal, persistent and pervasive level of inattention, hyperactivity, or impulsivity that are usually apparent before 7 years of age, that impair school performance, interpersonal relationships, are often co-morbid with other conditions, and they can lead to academic failure, increased risk for delinquency, increased risk for accidents, car accidents, in particular, and increased overall medical cost. This does not mean that all children with Attention Deficit Disorder are doomed to have a life of failure, but on average as a group, children who have Attention Deficit Disorder are at greater risk for this negative outcome than peers who don't have Attention Deficit Disorder.

Attention Deficit Disorder is a clinical diagnosis, even if there is a biological substrate for it, of course, but there is no diagnostic marker that can be used for making the diagnosis. So it's a sort of labor-intensive sort of time consuming diagnosis that requires careful documentation and examination of a patient, collection of information from the patient, the parent, the teacher. It requires a comprehensive assessment to rule out other possible causes of symptoms. It also includes a comprehensive physical and neurological examination.

It is estimated that at least 5 percent of school-age children meet the criteria for Attention Deficit Disorder. A higher rate, however, has been reported, so some people believe that the true estimate is more about 6-7, rather than 5. It's more common in boys.

Now there are different treatment for Attention Deficit Disorder, so there are choices. For instance, there is behavioral therapy which is an effective treatment of Attention Deficit Disorder, and then there are a number of pharmacological agents, of which amphetamines, methylphenidate and atomoxetine are approved officially by the Food and Drug Administration. Other drugs, like bupropion, Modafinil, Clonidine or Guanfacine, or Tricyclic Antidepressant have some evidence of efficacies, in some cases also good evidence, even though it's less strong stimulants for sure, but they do not have an indication at the moment for the treatment of Attention Deficit Disorder.

Stimulants are drugs, as has been already pointed out in the previous presentation, that stimulate the central nervous system, they increase alertness, attention, they decrease fatigue, they decrease and focus motor activity; and, therefore, they decrease hyperactivity. They increase goal-oriented activities, improve performance in a variety of tasks, decrease also appetite and sleep, and can cause euphoria; and, therefore, can be abused.

The efficacy of stimulants in the treatment of Attention Deficit Disorder is very well documented by numerous placebo-controlled studies using both crossover, which is within subject the design, or parallel-group design, and this is true for both methylphenidates and a variety of amphetamine products. The efficacy has also been confirmed by several meta-analyses, so in some way in evidence-based medicine, the strength of the evidence is usually ranked from 1-5, where one is the highest, and five is the weakest. And the efficacy of stimulants in Attention Deficit Disorder certainly meets the highest standard for evidence. Strong evidence from at least one systematic review of multiple well-designed randomized clinical trials.

In this particular study that was funded by the National Institute of Mental Health, and also the Department of Education and other federal agencies, more than 200 children age 7-9 suffering from Attention Deficit Disorder Hyperactive and Inattentive type combined, were randomized to receive a sequence. Each subject received placebo in three different dosage of methylphenidate, and the teacher and the parent scored blindly the symptoms of Attention Deficit Disorder. The higher the score here, the more is the severity of attention deficit symptoms. You can see there is a dose effect curve that is evidenced particularly with the teacher scores, but also with the parents. To say that of all the children who entered into the study, four subjects did not tolerate methylphenidate and they had intolerable side effects; and, therefore, they had to be discontinued and were not included in this analysis. It's about 2 percent of the patients, so the rate of intolerable side effects in this large sample was 2 percent.

In this particular study of -- so, yes, it can characterized with effect size which is the magnitude of effect of methylphenidate versus placebo expressed in standard deviation units is 1.3 if we use the teacher scores, and it is .6 if we use the parent. So the fact is certainly moderate and large actually, a large effect size if you use the teacher. Response rate, 77 percent of the children were responders to methylphenidate, and 12.5 percent on placebo. The remaining 10 percent didn't respond to either, so the non-response to methylphenidate is 10 percent roughly, and 12.5 response also on placebo.

The number needed to treat in order to add one extra patient who improves to the one that would improve on the control condition, in this case placebo, is about 1.5, which is one of the highest in medicine, in general, not just in psychiatry, but in medicine. If you consider that the ceiling is one, because you cannot go better than that, so it's really at the top in terms of discriminating between placebo and active medications.

The same conclusion you can reach by using parallel-group design. The previous data were within subject design, this is parallel-group design, this is study supported by industry where about 300 children were randomized to receive placebo or methylphenidate, and the effect size using teacher ratings is .8, a large effect size, and .9 using the parents scores. And, therefore, the number needed to treat in this particular group is 3.3. It varies somewhat because of a definition of improvement changes somewhat from one study to another, but still the number needed to treat is very favorable.

If we look at amphetamines, we pretty much get the same result. And, in fact, there is no evidence that methylphenidate is any different to amphetamines with efficacy, so they are considered to be sort of equivalent in terms of magnitude of efficacy. So in this particular study, which is a parallel-group design, short term about three weeks, about 60 patients were randomized to amphetamines, mixed amphetamine salts or methylphenidate, and the primary outcome were the teacher ratings. And you can see, this is a measure of inattention and over-activity. This is the placebo. This is the methylphenidate and this is the amphetamine, and clearly the amphetamine, and also the methylphenidate are better than placebo. This is course of aggression and defiance that are not really part strictly speaking of Attention Deficit Disorder, but they are often associated with Attention Deficit Disorder, and there is an effect on that, too, even though somewhat smaller. These are teacher ratings, and these are the parent ratings.

In this other study, also very large study, also supported by industry, about 600 children are using a parallel-group design, were randomized to receive fixed doses of amphetamines, 10, 20 and 30 are placebo, and teachers and parents rated the symptoms. And you can see that this is a level of placebo, and there is also a dose response curve relationship with the linear with amphetamine, as we have seen also with methylphenidate.

The results in adults are less extensive, because there are much fewer studies in adults, both with methylphenidate, with amphetamine, than with children. However, the conclusion seems to be the same. In this study in adults, about 150 patients with Attention Deficit Disorder, parallel-group design, placebo controlled, effect size 1.4. The response rate was 76 on methylphenidate, and 19 percent on placebo, number needed to treat 1.4.

For amphetamines, smaller sample size in this case, a response rate 70 percent on amphetamines, 7 percent on placebo, number needed to treat 1.6. So the signal is detected in adults as it is in children, and with the same magnitude.

The therapeutic benefit emerges quickly basically 20 minutes after the first dosing persist as long as the medication is administered and disappears if the medication is discontinued. Basically, all children, even if they're treated extensively, when they are switched to placebo they represent the same symptomatology of Attention Deficit Disorder. Stimulants, they suppress symptoms, of course, may not eliminate them for good.

Atomoxetine is another drug that has proven efficacy in Attention Deficit Disorder. In this study that was supported by the maker of atomoxetine, Eli Lilly, 171 children were randomized to a parallel-group design versus placebo, and the outcome were Attention Deficit Disorder symptom rated by the clinician, the parent, and the teacher, effect size 0.7, response rate 59 percent on atomoxetine, and 31 percent on placebo, number needed to treat 3.6.

This other study, about 300 adults, parallel design, goes up to 60 milligram a day, outcomes, rate and scales by the clinician this time, effect size more modest. This is a small effect size, still statistically significant.

There is an interesting study that is not published yet that was presented last year at a Psycho Pharmacology Conference sponsored by NIMH, which directly compare atomoxetine with a preparation of methylphenidate in about 500 children with Attention Deficit Disorder, which also included a placebo group. And the doses were I think a fair comparison, a fair amount of methylphenidate, and a good dose also of atomoxetine, which is quite important when you do this type of comparison. The response rate of 56 percent on methylphenidate, or in this case atomoxetine 45 percent, the placebo 24 percent, OROS was statistically significant better for atomoxetine. And this study was supported by Eli Lilly, which is the maker of atomoxetine.

The multi-modal treatment of Attention Deficit Disorder was a study conducted in the 90s, and was publicly funded by NIMH, Department of Education, NIDA, other federal agencies. And it's remarkable because you compare different treatment strategies for the treatment of Attention Deficit Disorder in children age 7-9. And the conclusion was that careful use of medication - one of the conclusions, careful use of medication, I would say intensive use of medical, so at fairly high doses, is more effective than behavioral therapy. In fact, the number of children with events over months of a study, so it was not short-term, was long-term, 14 months, randomized study that were optimally controlled in their Attention Deficit Disorder symptoms was 56 percent on medication, and 34 percent on behavioral therapy.

These are some results of MTA study. For instance, the fact on an attention score rated by the parent or by the teacher, this is time as expressed in day, but this is the last point, is 14 months basically, so a little longer than one year. These are the symptoms, and so the lower the score, the better the trial with respect to the outcome. You can see that there's two conditions that are medication management or medication plus behavioral therapy in combination are better than behavioral therapy alone according to the parent, and according to the teacher.

If we look at this other outcome, which is hyperactivity rated by the parent, but the same result you will get from a teacher, you get the same response, the same result; meaning that the medication, either alone or in combination with behavioral therapy is better than behavioral therapy alone or the community control. That was the comparison in this study.

And interesting enough, also, the social scales of a child as rated by the teacher improve. In this scale, the higher the score the better, of course, so there is an improvement in social skills that is higher on the medication compared with the other conditions.

And also, I think even more interesting, if you look at other symptoms that have nothing to do with Attention Deficit Disorder, so-called internalizing symptoms such as anxiety, moodiness, self-esteem, they tended to improve, meaning the symptoms tend to decrease with the medication, so the efficacy of the medication is just not limited to the core symptoms of Attention Deficit Disorder, but extend and apply to a variety of other symptoms that the children have.

In conclusion, there is strong evidence that pharmacotherapy with the drugs that we have considered improve a patient with Attention Deficit Disorder with respect to symptoms of hyperactivity, impulsiveness and inattention, which are the core symptoms of Attention Deficit Disorder. And also, there is an effect on defiance and aggression, that are often associated with Attention Deficit Disorder. There is an improvement in interpersonal interaction and social skills, in general. The results are an effect on academic work in terms of amount of academic work that is done, completeness, and accuracy. However, an effect on academic achievement such as school grades, has not been demonstrated. Actually, that was one of a few outcomes, probably the only outcome that was included in the MTA, Multi-modal Treatment of Attention Deficit Disorder, where there was no treatment effect. This is maybe due to the fact the school grades have such large variance that is not really amenable to pick up a treatment effect, so it could methodological effect. But the reality is that still we don't know whether successful treatment of symptoms of Attention Deficit Disorder translate into a better outcome, a better prognosis for the child as the child becomes an adult. So we still don't know if treatment really is able to change the negative outcomes of Attention Deficit Disorder that will lead to academic under-achievement, occupational under-achievement, in general, higher rate of accident, and higher medical cost. And since I'm involved in research and I like to divide research in different chapters, I would say the first task which was to show the effectiveness of treatment of a short-term base was achieved back in the 70s and 80s, and well documented later. But the treatment is effective in the long-term with chronic use. This was very well documented in the 90s, and MTA played a major role in that.

The challenge now is to try to design studies that can address this question: Does treatment lead to distal benefits? And again, this is the challenge, I think, for researchers right now. Thank you.

DR. NELSON: Thank you, Ben. The next presentation is by Susan McCune. Susan is a neonatalogist who is in the Division of Pediatric Drug Development at the FDA.

DR. McCUNE: Thank you, Dr. Nelson. Good morning, ladies and gentlemen of the Committee. Thank you. I am going to present today the one-year post-exclusivity adverse event review for Adderall XR.

This slide gives you an overview of my talk. First, I'm going to give you a little background information about Adderall XR, but then I'm going to briefly describe the clinical trials for the initial approval in adults and children 6-12 years of age, and then again briefly the clinical trials that were done for exclusivity in the adolescent population. I'm then going to switch gears and talk about the Adderall XR use information, and finally, we're going to talk in some detail about the adverse event reports for Adderall XR in the one-year post-exclusivity period, which covers October 28^th, 2004 through November 28^th, 2005.

In terms of background drug information for Adderall XR, Adderall XR capsule contain dextroamphetamine saccharate, amphetamine aspartate monohydrate, dextroamphetamine sulfate, and amphetamine sulfate. The therapeutic category is that of central nervous system stimulant. The sponsor is Schier. The indication for the use of Adderall XR is treatment of Attention Deficit Hyperactivity Disorder, or ADHD. The original market approval for Adderall XR was October 11^th, 2001.

Of note, Adderall was originally marketed in 1960 under the trade-name Obetrol. It was then approved to treat ADHD and narcolepsy in 1996. Pediatric exclusivity for Adderall XR was granted on October 28^th, 2004.

In terms of some background information on Adderall XR, you've heard some of this information already from Dr. Vitiello, and Dr. Andreason. The mechanism of action of Adderall XR in ADHD is unknown; although, like methylphenidate it is thought to block the reuptake of both norepinephrine and dopamine, and the diagram here shows only dopamine, but should also show norepinephrine, blocks the reuptake of norepinephrine and dopamine into the presynaptic neuron, and increase the release of these monoamines into the extraneuronal space.

In addition, there are known peripheral actions that include elevations of both diastolic and systolic blood pressure, and weak bronchodialator and respiratory stimulant actions. In terms of the drug delivery system for Adderall XR, each capsule contains two types of drug-containing beads designed to give a double pulse delivery of amphetamines allowing for once daily administration.

If you look at the drugs to treat ADHD, and once again, you've heard this this morning from a number of speakers, but I'd like to put them into categories for you. This is based on the clinical practice guidelines that were published in 2001 in pediatrics. In addition to the drugs that you see here, we need to add dex methylphenidate which is phocalin into the methylphenidate category. And as you've heard this morning, Amodafinil is an additional drug. So in terms of the stimulants which are identified as first-line treatment, there are the methylphenidates, there are the amphetamines, and as you heard this morning, Pemoline has been taken off the market. They are described as non-stimulants, the atomoxetine, and in addition, there are second-line treatments, including the antidepressants, the tricyclic antidepressants of bupropion, and these are not approved for ADHD treatment.

Today in my talk, I'm going to focus on Adderall XR. I'm also going to give you some information about Adderall, and just as a clarification, although it's classified here as intermediate acting, it would be considered a short-acting drug. In terms of looking at the literature and what are described as adverse events for Adderall XR, this is from Dr. Rappley's paper in The New England Journal from 2005. And the side effects of Adderall XR include appetite suppression, weight loss, stomachaches, headaches, irritability, possible growth inhibition, exacerbation of psychosis, exacerbation of tics, mild increase in blood pressure and pulse. And the drug is contraindicated in patients who have cardiovascular disease, hypertension, hyperthyroidism, glaucoma, drug dependence, or use of monomenoxidates inhibitors.

Now we're going to talk for just a brief minute about the clinical studies that were done for Adderall XR initial approval. These trials involve the double-blind randomized placebo-controlled parallel-group study of 255 adults. In addition, a double-blind randomized placebo-controlled parallel-group study of 584 children age 6-12 years. An additional classroom analog study of 51 children age 6-12 years was compared to placebo. In all of these studies, there was demonstration of significant improvement in patient behavior.

In looking at the exclusivity studies that were done for Adderall XR, the clinical trials in adolescents involved a pK trial of 17 adolescents who were less than 75 kilos, and 6 adolescents greater than 75 kilos. A double-blind randomized multi-center parallel-group placebo-controlled study of 327 adolescents age 13-17 years. There was a primary cohort of 287 patients who weighed less or equal to 75 kilos who were randomized to a fixed dose treatment for four weeks. This treatment included placebo, 10 milligrams, 20 milligrams, 30 milligrams, or 40 milligrams once daily of Adderall XR in the morning. A secondary cohort was also studied, including 40 patients who weighed greater than 75 kilos. And these patients were randomized to a fixed a treatment for four weeks of either placebo, 50 milligrams, or 60 milligrams once daily of Adderall XR.

The primary efficacy variable was the ADHD rating scale for total scores for the primary cohort. Improvements in the primary cohort were statistically significantly greater in all four primary cohort active treatment groups compared with placebo. However, there was not adequate evidence that doses greater than 20 milligrams per day conferred any additional benefit.

Following the studies for exclusivity, additional information was added to the label. The studies for pK were defined in the clinical pharmacology section of the label. The clinical studies were documented for adolescents in the clinical study section, and the dosage administration section was altered, stating that the recommended starting dose for adolescents who are 13-17 years of age who were diagnosed with ADHD is 10 milligrams per day. And the dose may be increased to 20 milligrams per day after one week if ADHD symptoms are not adequately controlled.

In terms of the clinical trials, discontinuation and adverse events for the exclusivity trials, there were eight patients, or 3.4 percent of the patients that were enrolled in the trials that discontinued treatment due to insomnia, depression, motor tics, headaches, light-headedness, and anxiety. The adverse events that were reported in 5 percent or more of patients in the exclusivity clinical trials were abdominal pain, loss of appetite, insomnia, nervousness, and weight loss. And 2-4 percent of patients reported accidental injury, asthenia, or fatigue, dry mouth, dyspepsia, emotional ability, nausea, somnolence, and vomiting.

Now I'm going to switch gears and talk about the drug use trends for Adderall XR in the period from 2003 to 2005. The number of prescriptions that were dispensed by retail pharmacies increased from approximately 7.3 million in the pre-exclusivity period - that would be November of 2003 to October 2004 - to approximately 8.6 million in the post-exclusivity period - November 2004 to October 2005. And this data comes from Verispan.

Psychiatrists were the most frequent prescribers at 31.8 percent, followed by pediatricians, who were the second most frequent prescribers at 29.7 percent in the post-exclusivity period. And pediatric retail pharmacy prescriptions - in other words, prescriptions for patients zero to sixteen years of age - accounted for 63.8 percent of all Adderall XR prescriptions in the post-exclusivity period.

I want to take just a minute to walk you through this slide, because you're going to hear this data, I think, again later on in the talks. Basically, this is a percentage of total prescriptions on the Y axis. Along the X axis you have three time periods, 2002-2003, 2003-2004, and 2004-2005. And the Office of Drug Safety examined eight therapeutic classes of treatments for ADHD. I'm showing you the top three here, which include amphetamine, dextroamphetamine, atomoxetine, and methylphenidate. There were five other therapeutic categories that were looked at that included dextroamphetamine, dex methylphenidate, methamphetamine, Modafinil and Pemoline, just to give you the total universe of what was examined. And as you can see, the amphetamine products accounted for approximately 35 percent of use over that time period, and was relatively constant. The use of methylphenidate decreased slightly over these three time periods, and the use of atomoxetine increased slightly over those three time periods.

In looking at the category of the amphetamine and a dextroamphetamine combinations that I showed you previously - now this is just a breakdown within this amphetamine, dextroamphetamine combination group - once again, the number of total prescriptions on the Y axis, and the same three time periods on the X axis, and you can see that the majority of use in the amphetamine and dextroamphetamine combinations is accounted for by Adderall XR, a smaller amount by amphetamine salt, and less so by immediate release Adderall.

Now I'm going to focus on the pediatric adverse event reports for Adderall in the one-year post-exclusivity period. In terms of all of the reports, these are raw reports in the one-year post-exclusivity period from October 28^th, 2004 to November 28^th, 2005. There were 210 total number of reports. There were 98 reports in the pediatric population. Please remember that number because I'm going to walk you through that number in a minute, and there were eight deaths and I'll walk you through those numbers, as well.

For the next few slides, I'm going to give you information about both Adderall XR, as well as immediate release Adderall. I will refer to Adderall XR always as Adderall XR. I will refer to Adderall Immediate Release as either Immediate Release Adderall or just Adderall.

In terms of the Adderall XR review, there were 98 reports that I showed you in the previous slide. When we did a hands-on review of those reports, it was found that actually 29 of these were related to the Immediate Release Adderall product, and not to Adderall XR. And so, in going back, the Office of Drug Safety then did a subsequent review for Immediate Release Adderall to make sure that we were capturing all of the Adderall XR reports, as well. So in looking at the Immediate Release Adderall review, those 29 Immediate Release reports were also captured, of which two were adults. But in addition, there were three additional Adderall XR reports that were identified in the Adderall review, so if you look at just the Immediate Release Adderall, there were a total of 29 Immediate Release reports, two of which were adult, leaving us with a total of 27 total pediatric Immediate Release Adderall reports in the one-year post-exclusivity period.

Now if we go back to Adderall XR, we started with these 98 reports. We subtract the 29 that were related to Adderall, leaving us with 69 reports. Of those 69 reports, four were adults, 16 were duplicates, and four had no adverse effect that was noted. Please remember this four, because this will show up again in one of my subsequent slides. So this leaves us with 45 total reports. If you add the three from the Adderall Immediate Release review, this gives us a total of 48 pediatric Adderall XR reports in the one-year post-exclusivity period.

In the next few slides, I'm going to give you data for both Adderall XR, as well as Immediate Release Adderall. These are the demographics in the one-year post-exclusivity period. For both Adderall XR and Adderall, the majority of reports were in the male population. The majority of patients used the drugs for ADHD or ADD, and the predominant age in which adverse events were reported were primarily in the 6-11 age range, and the 12-16 age range.

In terms of the outcomes, you've heard a little bit from Dr. Iyasu this morning that on the MedWatch form, the outcome boxes are checked. Please note that multiple boxes may be checked on one report, which is why these numbers don't necessarily add up to the totals of 48 and 27. So in the hands-on review, with Adderall XR there were five deaths, there were two deaths with Immediate Release Adderall, 14 patients required hospitalization with Adderall XR, six with Adderall. Eight patients reported life-threatening illnesses with Adderall XR, one with Adderall. Two patients reported disability with Adderall XR, and seven patients with Adderall XR required intervention, and two patients with Immediate Release Adderall.

Now I'm going to focus on those seven deaths, the five in the Adderall XR, and the two in the Immediate Release Adderall in that one-year post-exclusivity period. The first was a 10-year old who collapsed at home after 22 months of treatment with Adderall XR, 15 milligrams per day. The autopsy revealed coronary artery anomalies. Other family members were subsequently found to have short QT syndrome.

There was a 10-year old who experienced sudden death while taking Adderall XR, no other details were available in the report. There was a 12-year old who took methylphenidate for four years, who died suddenly after running cross country on the first day of Adderall XR treatment at 10 milligrams per day. There was no autopsy performed. The mother had a history of ventricular tachycardia. An 11-year old took Adderall XR, 15 milligrams per day for approximately four months. Two months after discontinuing Adderall XR, the patient was found unresponsive and could not be revived. Autopsy listed cardiopulmonary arrest of obscure causes. At the time of death, the patient was on atomoxetine and bupropion. Then finally, a 14-year old male who made unusual movements and collapsed at school, he developed ventricular fibrillation, was subsequently hospitalized on full support, but died ten days later. No autopsy was performed. He had been on Adderall XR for three years.

I just want you to note for future presentations by Dr. Gelperin this afternoon, that she will talk about four deaths in the year 2005, and that's because this patient had been off Adderall XR for two months, so her analysis includes these other four patients.

In terms of the Immediate Release Adderall, in the post-exclusivity period there were two patients who died. One was a 12-year old who experienced sudden cardiac death while running. The patient had been taking Adderall, 30 milligrams per day for five months. Concomitant medication was atomoxetine. An autopsy found an unspecified genetic cardiac problem. And the second case was a 7-year old who died during sleep after restarting Adderall following a summer break. That patient at autopsy was found to have a bicuspid aortic valve.

In trying to put some of the deaths into a little perspective for you, if we go back to January of 1999 through the end of the post-exclusivity period of November 2005, we identified eleven sudden deaths of unknown ideology, six sudden deaths with either cardiac or genetic predisposition, four suicides, and two other. I want to try to put this in a little perspective because you all have received a number of reviews, and sometimes the numbers overlap, and it's difficult to identify where all these patients are coming from.

If you look at this timeline from January 1^st, 1999 through November 28^th, 2005, there was a review by Dr. Gelperin between January 1^st, `99 and December 31^st, 2003 that identified twelve sudden deaths. And then I have just told you about the seven sudden deaths that we identified in the post-exclusivity period. However, when you look at each of these reports, two of these reports in the post-exclusivity period were actually duplicate reports from the previous review, so that means that there actually were a total of 17, as I presented in the previous slide, 17 sudden deaths. And in the period of Dr. Gelperin's review, there were three suicides noted, there were no suicides that I reported to you in the one-year post exclusivity period; however, there was one suicide in this gap of time between December 31^st, 2003 and October 28^th, 2004. And of note, this was when Health Canada took Adderall XR off the market between February 9^th, 2005 and August 24^th, 2005.

Okay. In looking at the categories of adverse events reported for Adderall XR and Immediate Release Adderall in the one-year post-exclusivity period, there are 52 here, and that's because that includes these four patients lack of effect, and no adverse effect that I told you about before. So with the Adderall XR, there were 17 psychiatric adverse events, four psychiatric adverse events with Immediate Release Adderall. There were 15 cardiovascular adverse events with Adderall XR, and 6 with Adderall. There were five neurologic adverse events with Adderall XR, six with Adderall. There were three abnormalities in growth and nutrition with Adderall XR, and none with Adderall. There were two dermatologic adverse events each with Adderall XR and Immediate Release Adderall. There were three adverse events associated with the respiratory tract with Adderall XR, and then lack of effect, one with Adderall XR, seven with Immediate Release Adderall. Three other adverse events with Adderall XR, two with Adderall, and no adverse event noted in three cases with Adderall XR.

Now I'm going to focus solely on the Adderall XR adverse events that I just told you about in that previous slide. And before I go into any of the rest of these slides, I want you to note that each case may report more than one adverse event, so these numbers here do not necessarily add up to the number here. And just so you know, underlined events are those that are unlabeled. So in terms of the 15 cardiovascular adverse events, there were five cases of tachycardia, four chest pain, one prolonged QTC, three arrhythmias, four sudden death, one autonomic dysfunction, and two unspecified cardiac disorders.

In terms of the 17 psychiatric adverse events, there were four that described hallucination, seven that described violent behavior that included aggression, agitation, hostility, homicidal ideation or assault, three that described psychosis, four suicidal ideation or attempt, four crying, moodiness and irritability, two paranoia, three insomnia, one nightmares or night terrors, one depression, and two panic anxiety.

In terms of the five neurologic adverse events, there were two dyskinesias, one seizure, one depressed consciousness, one leg spasticity, and two tics. In terms of the respiratory adverse events, there was one patient who had a respiratory arrest associated with severe asthma, one patient that described dyspnea, throat tightness, and one patient with asthma. In terms of the dermatologic adverse events, one patient described toxic epidermal necrolysis, and one patient described generalized rash and skin exfoliation. In terms of growth and nutrition, one patient described growth suppression, two patients described weight loss, and there was one case of anorexia. In terms of the other adverse event reports, there were three; one suspected glaucoma, one increased prothrombin time, and one with extremely high amphetamine levels.

So in summary, for the Adderall XR adverse event report profile in the one-year post-exclusivity, there were 48 unduplicated pediatric reports. There were 14 cases of cardiovascular adverse events, including sudden death. These are known and labeled, especially in children with structural cardiac abnormalities, as Dr. Andreason told you about earlier this morning, and there are plans for future studies that will be discussed later today by Dr. Graham.

In terms of the 15 psychiatric events, these will be discussed in context at the other drugs used to treat ADHD later today by both Dr. Gelperin and Dr. Mosholder. And in terms of the serious skin reactions, these are not currently in the Adderall XR label. However, these same skin reactions that were described here are listed under the adverse events with other methylphenidate products in the Concerta label. And there was one case of glaucoma, and glaucoma is labeled as a contraindication.

Obviously, this report couldn't have been done without a lot of help from a lot of other people, both in the Office of Drug Safety and the Division of Psychiatric Products, and in the Division of Pediatric Drug Development. So that's the end of my talk, and now it is my pleasure to introduce Dr. Andrew Mosholder, who is a Child Psychiatrist and Medical Officer in the Division of Drug Risk Evaluation in the Office of Drug Safety.

DR. MOSHOLDER: Good morning, and actually, this is the first of two talks I'll be presenting this morning. And what I'm going to try to do to provide some context for the discussion is provide an overview of the prevalence of use of ADHD medications, and of ADHD in the United States.

Okay. I'll be covering, first of all, some recent survey data from the CDC that was published last September, and also some data that we have from the Verispan prescription drug database. And just for those of you who were at the February DsaRM meeting, this is very similar to the presentation that I gave there.

Okay. To go first to the CDC survey, this was part of the National Survey of Children's Health. It was a telephone survey conducted in 2003 and early 2004. There were roughly 100,000 subjects in the sample ages 4-17 years, and for each subject the parent or guardian in the household responded by telephone to questions about diagnosis and treatment of ADHD for that child. And the sample was such that they could do statistical projections to the national and the state level, which I'll be showing you in a minute.

All right. This is a fairly rich slide, so let me describe it, take a little bit of time to describe it for you. First of all, we have on the left males, and on the right females. Going along the horizontal axis out from the center is the percentages, and then the vertical axis is age. Now you see two sets of bars. The outer paler blue bars are the percent of subjects who have ever been diagnosed with ADHD, and the inner darker bars are the percent of subjects who are currently receiving a medication for ADHD at the time of this survey. Now you see that they were able to put statistical confidence intervals on these estimates, so several points to make here.

First of all, looking at the outer bars which represent the proportion or the percentage of subjects diagnosed with a disorder, you can see that in both girls and boys it increases steadily until say around age 9 or 10, and then tends to level off. And also, you'll see, it's very obvious that the diagnosis is more common in males than in females.

By the teenage years, there's some statistical variation, but one interpretation is that there are relatively few new diagnoses that are taking place in these older years. And if you average males and females together, that roughly 10 percent of children by the end of the teenage years will have been diagnosed with ADHD.

Then looking at the medication, which is the inner bars, you can see that again the prevalence increases until roughly around 10, 11, 12, both boys and girls, and peaks around 9 percent for boys in that age range, and around just under 4 percent for females. And you can see also that the prevalence of medication tends to decline a bit in the teenage years so that there are by the older teen years a fair proportion of the patients who are no longer receiving the medication.

All right. This map displays by state the prevalence of the diagnosis, lifetime diagnosis. You can see that there's quite a range from 6 to roughly 10 percent, and also you can see there appears to be some regional variation with sort of a higher prevalence in the southeast, and lower prevalences here in the west.

Now this is a similar display, but here we're looking at the prevalence of ADHD medication use. And again, you can see there's a good range here from around 2 percent up to over 6 percent. Again, a concentration of more or less here in the southeast, and relatively less use in the west. And the reasons for these regional variations are not well understood. So the conclusions from the survey, first of all, there's a high prevalence of ADHD diagnosis and medication use in children and adolescents. The estimate is 2.5 million children in this age range receiving medication for ADHD, and that is around 4.3 percent of all children in that age group. And, of course, it's higher for males than females.

The authors said that with this high prevalence of medication use, it's important to get a better understanding of both the benefits and the risks of medication treatment for this disorder. And, obviously, the implication here - any risks that we discover about these drugs would have potentially great Public Health impact given the high prevalence of their use.

There are regional variations in both diagnosis and medication use, as I said, and medication use peaks around ages 9-12, with peaks being 9.3 percent of boys age 12, 3.7 percent of girls at age 11. Now there are limitations to this. Of course, it's survey data, and has the usual limitations of such data. And also, it's dependent upon parental recall, but nonetheless, I think it gives us some insights.

Okay. I want to turn now to the Verispan data. This is prescription data from the Vector One National Source, and this is a database that includes data on prescription activity from retail pharmacies from various sources, data is available on prescriber specialty, patient age and gender, and includes roughly 1.8 billion prescriptions annually for 150 million patients, so you can see it's a broad sample. Limitations, does not provide data on indication or duration of treatment.

And the drug products are active ingredients we looked at, amphetamine, dextroamphetamine combination, which as you've heard as Adderall and Adderall XR, atomoxetine or Strattera, the D-Ismer of amphetamine which is Dexetrin, the D-Ismer of methylphenidate which is Focalin and Focalin XR, methamphetamine which actually is a prescription drug product but is very rarely used, as you'll see in a moment. There are various methylphenidate products, and Modafinil, which as you've heard, is under consideration for this indication. Finally, Pemolin, which has since been taken off the market for petato toxicity, as has already been mentioned.

This displays by age group the proportion of the total of prescriptions for these drugs. This was the most recent six-month period for which we had data. And actually, zero to four it's not -- it rounds down to zero, but there is a very small amount of prescribing for that very young age group, somewhat more in ages 5-9, and about half ages 10-19. Ages 20-49, maybe a quarter, and then somewhat surprisingly, adults over age 50, up to 10 percent of the use. And although it's not a topic for this particular committee, at the February meeting, there was concern about this, because obviously for cardiovascular risks this would be a population where that would be of significant concern.

All right. This shows the total prescriptions separated by age group, and we see pediatric age group at the top here. This is the adult use, and then there's a small amount of prescriptions which have unspecified age. And this covers roughly a three-year period.

A few things to notice. First of all, you can see that for both adults and children, use is increasing. Looking at the most recent data, you can sort of see visually that about one-third of the total is now used by adults. And this is per quarter, so this is roughly 2 million prescriptions per month for children, and about one million per month for adults. And also you see there's a seasonal variation for the pediatric use, and as people probably guessed, these nadirs correspond to the summer months. That's not seen in the adult data.

Okay. This is a display of prescriber specialties. Unfortunately, this is for all ages. I didn't have it separated out by just pediatric age group, but you can see, as has already been mentioned, psychiatry and pediatrics, the largest portions, just under a third each, some prescribing by other primary care specialties, and then a variety of other specialties with smaller percentages. And we can imagine that if we had limited it to the pediatric age group, that this would probably have been a larger percentage. And this is data which Dr. McCune has already showed you in part. This is the total prescriptions by active ingredient, and you can see this is the total up here for all prescriptions. And as already been described, it's basically three primary drugs being prescribed, methylphenidate the leading one, followed by amphetamine represented by Adderall, and this is atomoxetine. Interestingly, you can see that there's a seasonal variation for the first two, but not in the case of atomoxetine. And then the others, very small percentage of the total.

And this is just another way of looking at more or less the same data for the most recent six month period, where it was available, of pediatric use by active ingredient. Again, you can see methylphenidate is the leading one, amphetamine, d-amphetamine representing Adderall, and atomoxetine with about 16 percent, and the others of miscellaneous compounds with smaller percentages.

So the conclusions from the prescription data, there's been increasing use of drugs by both adults and children. As I mentioned, there's roughly one million prescriptions monthly for adults, and two million for children. If you think back to the CDC estimate of 2.5 million children currently receiving medication, that's in the same ballpark as this, if you figure one prescription per child per month. It works out to be about two million. Methylphenidates most frequently prescribed followed by amphetamine, and then atomoxetine.

And in closing, I'd like to acknowledge the investigators from the CDC for allowing me to share their data, and also in our Office of Drug Safety, Carol Pamer for her assistance with the Verispan data. And now it's my pleasure to introduce Dr. Kate Gelperin, a colleague from the Division of Drug Risk Evaluation, and she will be talking about the cardiovascular adverse events.

DR. GELPERIN: Good morning. This morning I'm going to update you on the cardiovascular risk of drugs used in ADHD with a focus on the pediatric population. I'm going to review the rationale for our safety concerns, talk about MedWatch reports, including sudden death in children. I'm going to discuss some calculated reporting rates, and some information about background incidence of these events, and then briefly just touch on non-fatal cardiovascular events.

As Dr. Andreason reviewed earlier today, the amphetamine and methylphenidate are adrenergic agonists, and it is known pharmacologically that increased adrenergic tone can be associated with ventricular arrhythmias and sudden death in some patients. There is a known effect of sympathomometic drugs on blood pressure, and that's described in some labeling. Dr. Andreason has previously described the labeling about cardiovascular risk factors, such as structural heart abnormalities.

The relationship between blood pressure and the risk of adverse cardiovascular outcomes is continuous, consistent, and independent of other risk factors based on data from observational studies in adults. The risks in children are not known.

Atomoxetine, the brand name is Strattera, is a selective norepinephrine reuptake inhibitor so it's not strictly speaking classified as one of the stimulants, but I would just call your attention to the current approved labeling for this product, which does include a precaution about effects on blood pressure and heart rate, and does describe in the placebo-controlled clinical trials in children that there was an increase of about 1.5 millimeters of mercury in systolic and diastolic blood pressure compared to placebo.

There are two publications I'm aware of in children using ambulatory blood pressure monitoring methods. One more recent one by Samuels, Franco and Wan was published this year which describes 13 subjects with a mean age of 12-1/2 years, who underwent ambulatory blood pressure monitoring, both on their usual stimulant medication, and placebo, using a placebo-controlled double-blind randomized crossover design. There were nine males and two females, six of them were on methylphenidate, four were on amphetamine, and one was on dextroamphetamine. Eleven patients had paired ambulatory blood pressure monitoring studies that were considered to be of sufficient quality for a full analysis.

After a three-day run-in followed by a 24-hour monitoring period, the subjects crossed over to the alternate therapy repeated ambulatory blood pressure monitoring. The subjects demonstrated elevations in both hemodynamic parameters derived from ambulatory pressure monitoring on active drug. Total diastolic blood pressure and waking diastolic blood pressure were significantly higher on active treatment. Total heart rate was significantly higher on active treatment. The authors point out that the rate pressure product, which is the product of systolic blood pressure and heart rate, which is an index of myocardial oxygen demand, was higher during active treatment, and the effects of this over long periods of time is just not known in children.

Very few long term studies randomized control double-blind trials have been done in children. Dr. Vitiello reviewed the multi-modal studies, and there is a Swedish study which also was a long-term study of amphetamine. However, these studies, and in addition, there are long-term extension trials which do not have a comparator arm in some cases, or in other cases they are open label, that do have some information about blood pressure. However, targeted cardiovascular outcome trials have not been done.

A review of MedWatch cases that was done for the PBCA review did suggest potential cardiovascular signal, but these data were not considered to be conclusive. The non-fatal cardiovascular reports include syncope, chest pain, myocardial infarction, stroke, arrhythmias. However, the cases are often not well documented. There have been sudden death reports, and I'll tell you a little bit more about these.

Last June there was a Pediatric Advisory Committee discussion about these cardiovascular and psychiatric adverse effects, and the Committee considered that the FDA should pursue additional means to better characterize the cardiovascular risks for all drug products approved for ADHD.

I'm going to tell you about some calculated reporting rates from MedWatch reports, but I want to mention the limitations of reporting rates. As Dr. Iyasu mentioned earlier this morning, there is considered to be under-reporting of spontaneous reports in the United States; however, the extent of this is not known, but it results in a numerator that is not reliable.

In addition, the denominator is not reliable. It's based on projected drug utilization numbers, so this is not an incidence. And a comparison of reporting rates to a background incidence, or even between drugs, is really only a rough estimate. And, of course, with spontaneous reports there may be confounding, there may be other drugs on board, there may be pre-existing conditions. However, the presence of other drugs and pre-existing conditions does not necessarily rule out that the drug in question has contributed to an adverse effect, so the Office of Drug Safety has conducted searches of the adverse event reporting system, safety database, we call that AERS, and we used a definition of sudden death in our review that has been used by the World Health Organization, which is that death occurred immediately or within 24 hours of an acute collapse.

In our analysis, we excluded cases in which the death was caused by a multi-drug overdose, because we felt this would not be relevant to the usual therapeutic use of the drug. We also excluded all cases in which drug abuse was reported, and we excluded cases in which death was most likely caused by something else.

For a perspective, I'd like to call your attention to the published incidence rates for pediatric sudden death in the United States. A review of this topic was done in 1996 by Liberthson and published in The New England Journal. It includes data on 469 sudden deaths from nine studies of large populations. The rate of sudden death in these populations ranged from 1.3 to 8.5 per hundred thousand patients years, with males consistently out numbering females. In two-thirds of the cases, a specific cardiac cause was identified. Extrapolation of these data suggests that each year several thousand Americans under the age of 20 years die suddenly from cardiac disorders.

For ages 1-30 years, the most common cardiac causes of sudden death including myocarditis, hypertrophic cardiomyopathy, coronary artery disease, congenital coronary artery anomalies, conduction system abnormalities, mitral valve collapse and aortic dissection.

This slide was shown previously at the February Drug Safety Advisory Committee meeting, and is based on work that was done by my colleague, Dr. Lourdes Villalba, who is a safety reviewer in the Division of Psychiatric Products. These reporting rates cover the time period between 1992 and February 2005. You also received information in your background package for a consult that had been prepared by me prior to this consult with reporting rates for a five-year period. And you can see that the reporting rates per hundred thousand person years for this larger time period are actually comparable to the five-year period that we had previously looked at.

Of note, however, atomoxetine was not included in that earlier review, because that drug was only approved in November of 2002. So, for instance, in this slide there are only two years of data for atomoxetine included.

Although the calculated background rates that you see in this slide do not appear to exceed the published incidence rates based on death certificate reviews, this is not completely reassuring because the extent of under-reporting is not known. For instance, if FDA has received 10 percent of potentially relevant cases, then the reporting rate would be approximately equal to the published incidence rates. However, if FDA has received only 1 percent of the true number of cases, the reporting rate would exceed the published rate.

This slide describes to you some of the characteristics of the 13 cases of pediatric sudden death for that 12-year period in which amphetamine or dextroamphetamine was considered a suspect drug. In all 13 cases, the suspect drug was identified as either Adderall or Adderall XR. The ages of the patients ranged from 7-16 years, and they were all male. The dosage ranged as you see, and the duration of therapy ranged from one day to eight years. Autopsies were done in 11 cases, and the autopsy findings included things like aberrant origin of coronary artery, idiopathic hypertrophic subaortic stenosis, bicuspid aortic valve. In three cases, the coroner stated that there was an unexplained increase or a toxic amphetamine level in a child who had been taking apparently normal doses.

Cardiac hypertrophy was described in three cases. A heart murmur had been reported on history in three cases, a maternal history of ventricular arrhythmia was described in one case, and there were no relevant autopsy findings in five cases. Ten of these 13 cases, there were no concomitant medications reported.

With regard to the 11 cases identified of pediatric sudden death in that 12 year period in which methylphenidate was considered a suspect drug, the ages ranged from 9-15 years. There were seven males, and there were four females. The brand name of the product was Ritalin in seven cases, and Concerta in four cases. The doses ranged as you see. The duration of therapy ranged from two months to ten years, and an autopsy was done in seven cases.

In two cases, there was reported a congenital cardiac malformation along with concomitant quantity in therapy. In one case there were multiple abnormalities, including heart hypertrophy and tricuspid valve anomalies. In one case there was cardiac small vessel damage, cardiac hypertrophy and obesity, and in one case there were unexplained toxic levels of methylphenidate, and this was nine days post-operative. One subject had a history of syncope and there were no pertinent findings reported in five cases. In six of these cases there were no concomitant medications.

With regard to the three atomoxetine cases reported in that 12-year period, the youngest child was 2-1/2 years old, and the oldest was 12 years old, two were male and one was female. The duration of therapy ranged from six weeks to four months, and an autopsy was done in all three cases. One autopsy found toxic levels of a different drug, Olenzepine, and this may have caused the death. In one cases, lymphocytic myocarditis compatible with a viral infection was found at autopsy, and in one case peribronchioloar chronic inflammation was found. In none of these three cases were structural heart abnormalities noted. In two of the cases, there were no concomitant medications.

This slide attempts to update for you so that we can bring this information up to the most recent time possible, just for the calendar year of 2005, so these reporting rates are going to include two cases which were actually included in Dr. Villalba's reporting rates which did go up to February of 2005. The reason this was done is just to apply - we now use a different vendor which is Verispan, and so to use the drug use data that was available, the numerator was selected to be the year 2005. So for the drugs of interest, you can see the count of cases for methylphenidate - there were two, for amphetamine there were four, and for atomoxetine there were four, and the corresponding reporting rates we believe are probably not meaningfully different from the previous reporting rates. However, I don't mean to say that these are completely reassuring numbers.

I will just briefly describe the cases, although you really have already heard about these from Dr. McCune. And as she mentioned in the amphetamine reporting rate, the case where the child had stopped the amphetamine two months prior to death and was actually taking atomoxetine and Buproprion at the time of death was not included in this reporting rate because the drug had been stopped two months prior to death.

The ages ranged from 10-14 years, with a mean of 11.5 years. There were three male and one female. The brand was Adderall XR in three cases, and Adderall in one. The dose is 15 milligrams total daily dose in two cases, 30 milligrams in one case, and not reported in one case. Duration of therapy ranged from five months to three years. Autopsy was done in two cases. One the report stated that the coroner found a genetic cardiac problem which was not further specified. The other case, the finding was coronary artery anomalies. There were no concomitant medications in two cases.

For methylphenidate there was very little information with the two cases that were reported. And atomoxetine, the four cases ranged in age from 6-11 years. They were all male. The duration of therapy ranged from three days to a few months. All four had autopsies. One showed cardiomyopathy and valvular disease, one showed brain herniation, one stated cardiopulmonary arrest of obscure causes, and there were no concomitant medications in two reports.

This pediatric sudden death case report, I think, highlighted for us the risk of a child with previously undiagnosed severe structural heart abnormality starting therapy with a stimulant, and that this 13-year old male who had previously been healthy and involved in sports was started on a 20 milligram dose of Immediate Release amphetamine, and he did collapse at his computer in the late afternoon, and the autopsy showed idiopathic hypertrophic subaortic stenosis.

The non-fatal cardiovascular events are under review, but in a five year period for amphetamine, you can see they included things like syncope, arrhythmia, myocardial infarction. For methylphenidate, similar types of events, and for atomoxetine, these events are currently under review.

In summary, I would just state that we have many challenges in risk assessment of this issue, the acute versus chronic effects of the drugs, the different background risk for the different age groups, and the unknown effect of confounders. I'd like to introduce to you Dr. David Graham, who is our Associate Director for Medicine and Science, and he's going to tell you about the FDA plans to study these issues.

DR. GRAHAM: Good morning. I will spend the next few minutes talking about studies that we have in mind to investigate further the question of cardiovascular risk with drugs used to treat Attention Deficit Disorder. Just sort of summarizing what's gone before, within the Office of Drug Safety, I think it's safe to say our Bayesian priors are that these drugs probably do increase cardiovascular risk. The question then really is, by how much, and are there factors that could predict who might be more likely to have this happen?

The distinction in our mind between a stimulant and a non-stimulant is, in our view, probably immaterial to the question of cardiovascular risk, because all drugs increase norepinephrine levels within the synapse. And we know very well, and if you look at any cardiovascular textbook, increased adrenergic tone is associated with arrhythmias and sudden death, and so we think that all three of these drugs should be considered together as a group, and the distinctions I think are more a red herring than anything else.

Just to give you a little background, we have an epidemiology contract program that replaces a previous cooperative agreement program we've had that enables us to do epidemiologic studies in population-based settings. So these aren't clinical trials, they're observational studies. They use automated healthcare data, but we do have the ability to go back to medical records.

We currently have four programs that are funded covering about 23 million covered lives. One of the problems with these types of databases are turnover, and so in terms of trying to look at chronic use of a therapy over time, the turnover within these health plans gets in the way of our being able to follow people because they disappear from the health plan. An employer changes the health insurance for the family and they disappear from the data set, and so you can see the turnover varies at one year between 8-30 percent depending on which of the various plans we're talking about. And then for the five-year turnover it's 25-80 percent. This year we have funding of $1.6 million for databases. Next year that funding drops to less than a million dollars.

To investigate this question, we queried our four databases to see was it possible to go forward with a more in-depth formal epidemiologic study. The design that we employed was in section cohorts, so these are new users of these products, so there's a period of time that we see before they get their first prescription. And you can see the study periods that we've included, and the drugs of interest are the three major ADHD drugs that we're considering. And we'll be looking at two components, children and adolescents, and then adults, as well. And the rest of my talk will focus on the pediatric age group.

Our primary outcome of interest is sudden cardiac death, then acute myocardial infarction, cerebrovascular accidents, and arrhythmia are also important to us.

This just gives you a little detail about the population that we'll be hopefully doing the study in. We have a base population of about 7 million children under the age of 19, between 1-19 years of age, and that covers about 45 million person years of observation time. For each of the three drugs that we'll be studying, you can see the number of patients that we have, about 200,000 children on methylphenidate or amphetamine, and a smaller number with atomoxetine. And then the number of person years of time on drug that we have within each of these cohorts. So these are fairly large cohorts, but as you'll see, they may not be large enough to resolve uncertainties at the level we would like.

For all three drugs across the databases, the proportion of drug that's used in males is about is about 73 percent. This slide describes the persistency of use of these products over time, so this is a description of duration of use of products over time. And you can see for methylphenidate and amphetamine that the median duration of use is somewhere around seven or eight months within the database that we have, and that about 10 percent of patients have use that goes on beyond about 18 months. For atomoxetine, the curve is shifted towards shorter durations of use. That is, in some regards, an artifact of how long the drug has been on the market. But what it does tell you is that in terms of looking for risk as a function of time on drug, that the denominator at risk gets very small with prolonged durations of use.

Background rates for sudden death were summarized by Kate in her presentation. This is just to give the Committee an idea of what we think we're dealing with, which is that these are uncommon events, and so that poses unique challenges in trying to study something in an observational database. But it also means that the ability to study this in a clinical trial is probably nil. And you can see now that we have uncertainty about what these background rates are for sudden death between one and nine per hundred thousand per year. For myocardial infarction, the range is even greater depending on the study that's been done. And then for cerebrovascular accident, there we have a number of different studies that have all come up with about the same rate, so we think that in terms of certainty of background rates, our background rate for cerebrovascular accident is probably the one with the greatest certainty.

With this feasibility study we queried the database of these four health plans to see how many children they had with a primary diagnosis, hospitalized diagnosis with the different outcomes that we're interested in, so there's no replication of children in these slides. It was the first, the primary diagnosis. You can see that we had about 17 children with myocardial infarction, another 17 with other hospitalizations for ischemic heart disease, 14 with cardiac arrest, 49 with stroke, and 245 with arrhythmia. This is both superventricular and ventricular arrhythmias, but it was the primary discharge diagnosis for those hospitalizations.

If we look for any diagnosis within the hospital discharge, not only the primary but also secondary discharge diagnoses, we have 24 children with a diagnosis of myocardial infarction, about 90 with a diagnosis of stroke, and about 450 with a diagnosis of cardiac arrhythmia. Two hundred and forty one deaths were reported within this cohort, but as I'll show you in the next slide, that is a gross under-estimate of what the actual number of deaths probably were. First off, these deaths occurred at any time after entry into the inception cohort, and so you could have somebody who's entered the inception cohort, is on the drug for six months, and then is still in the database after that but isn't taking the drug any more, and they died subsequently, so we don't know if the deaths are while they're on drug or off drug. These deaths could be from any cause, not just from cardiac causes.

For two of the sites, the deaths that we had were based on in-hospital discharge diagnoses only. Or one site, we have no information on deaths, and for one other site we included both in-patient and out-patient deaths because they have linkage to death certificates. So when we go on to do our study, we will be looking for both in-hospital and out-of-hospital sudden deaths, and that will require that we go to the National Death Index, and that we seek out death certificates on all subjects that died.

The next five slides I'll go through quickly. I'll explain the first one, and then members of the Committee want to look at them at their own leisure, they can. But these are power curves that give a sense of what types of risk ratios we might be able to detect based on the background rates that we're dealing with, and the number of subjects that we have to work with in the study. And so what this slide shows is it's looking at acute myocardial infarction in children. And if you remember, the background rates there went between 1-20 per hundred thousand per year, so these power curves are predicated on a background rate of about 15 per hundred thousand per year. And what this slide shows is that if you go to the previous slide where we showed the size of the cohorts, we have a total of 400,000 children who are on any of these drugs. I mean, 400,000 person years, 399,000 person years on the drug, and so if you went over to the 400 and go up, you can see that we had a lot of power to detect even relative risk of two, if we look at all drugs combined together. But if we wanted to look at individual drugs, for example, we want to look at atomoxetine where we have about 40,000 patient years of use, you can see that we really only have power to detect a relative risk of somewhere around three or four. So this is just to give you an idea of what we're dealing with.

And this slide is for stroke in children where background rates were three of 100,000 per year. And then this is another way of calculating power, which is the way it's done in clinical trials, where a study is powered based on the number of outcome events that you'll experience. And so this, you can see, you can use this slide to see well, we need X number of events to determine the relative risk of a given magnitude.

Another way of looking at an observational study, or any study for that matter is - one way is what relative risk can we detect with what degree of confidence? Another way to approach it is what level of risk can we exclude with a given level of confidence, it's sort of basically capping the risk to say that we can't tell you exactly what the risk is, but we can tell you that it's unlikely to be greater than some number. And so the next two slides show what those power curves look like in terms of the power to exclude a variety of relative risks.

This slide tries to summarize what the previous five slides were dealing with in terms of what we believe the power is in our study to detect particular relative risks for sudden death, myocardial infarction, stroke in children by individual drug, and then combined. And so you can see that at the end of the day, there'll still be some level of uncertainty in what we can only hope for, probably, with the studies that we're proposing is, is that we'll detect, if they're there, relative risks that are relatively substantial, but the relative risk itself may be substantial, but the absolute risk may still remain relatively low.

Some additional power considerations I think that we need to be aware of is that we don't know what the background rate is for sure, especially say with myocardial infarction, with sudden death. And so based on the information we have, as I've circled in red here, it's possible that for myocardial infarction that we may have more cases than one would expect. We may have less cases than one would expect. And then again, we also have the problem, we don't know if these events happened while on drug or off drug, so these are things that need to be finalized in an in-depth study.

So I'd like to leave the Committee with a couple of caveats. One, this feasibility study we're talking about very preliminary results. The purpose of the preliminary feasibility study was to see whether there's enough information there for us to go forward with an in-depth study. We believe there is, and we're in the process of trying to design the study to go forward.

Definitions of exposure in outcome we used are relatively crude here, and they haven't been validated; although, acute myocardial infarction has been well validated in adults and the positive predictive value of a primary diagnosis of AMI is about 95 percent. Our power calculations are crude, and a lot of that has to do with the uncertainty about what background rates are.

At the February Drug Safety and Risk Management Advisory meeting, one proposal came up during the committee discussion about forming an observational echocardiographic study of patients who have been on these various drugs for varying durations of use. I think the committee was probably thinking more about doing such a study in adults than in children, but the same study design could be applied with children, as well.

Generally, what they were talking about was within a large healthcare database, identify patients on these drugs for varying durations of time, select suitable untreated patients, and perform echocardiography on them. And look to see what the left ventricular wall thickness is, what there contractility is, and then other measures that the echocardiographers could tell us would be important, perhaps predictors of various cardiovascular outcomes. So to give you an idea of what that might look like, taking our preliminary feasibility results and looking at the curve, the contour of duration of use, you could see at six months you'd sample people who were on the drug for six months at that particular time, bring them into a lab, do cardiography on them. You take people at 12 months, at 18 months, at two years, and do the same thing, and then what you basically have is sort of a time series study, not on a given individual being followed chronologically over time, but it's a cross sectional prevalence study over time of what echocardiographic changes might be.

So in summary, I think we have some concern about the potential for cardiovascular risk with these drugs. There's a very high prevalence of use of these drugs in children, and so the Public Health impact could actually be of importance and meaningful. Sudden unexplained death is our primary interest, and will also be the most difficult to study.

Our feasibility study has showed that we have substantial exposure time, and that the study is feasible to do. We have limited statistical power to detect relative risk below about 2, but the number of arrhythmia cases struck us as being particularly high. None of us being pediatric cardiologists, we didn't know whether that was something to worry about or not, but that was the primary discharge diagnosis. We're now in the process of obtaining cost estimates for carrying out the study, but our budget is very limited, and we're looking for ways to fund the study. And then these are the members of the study team from FDA and the various sites, research database sites that we've used to do the study. Thank you.

DR. MOSHOLDER: Hello, once again. And the next two talks we're going to be changing topics now, and we'll be talking about neuropsychiatric adverse events with drugs for ADHD. And I'll be starting with a presentation of clinical trial data, and then Dr. Gelperin will present post-marketing data. So for my talk, I'll describe the background of this project. You've already heard a bit about it this morning. And then I'll be focusing on clinical trial data regarding these events, and I'll have a few comments to make about some literature reports.

Okay. Just for orientation, this is a list of the currently approved products for ADHD, and I won't read all of them, but in general, you can see we have amphetamine, which as you've heard is Adderall, Adderall XR. There are a number of methylphenidate products, including extended release preparations, Concerta, Ritalin LA, atomoxetine, or Strattera. The D-ismer of methylphenidate represented by the Focalin products. There's a D-ismer of amphetamine which is Dexedrine, and then as you see, there's some other methylphenidate products. And then finally, there are two products for which this indication is pending, Modafinil, which will be discussed tomorrow, and also for methylphenidate, there's a methylphenidate transdermal system, a skin patch, for which we also had data. Although that's not marketed, so, of course, there's no post-marketing data.

Just to give you the background on this project, the BPCA review for Concerta last year, which was presented to this Committee, and a comparable review for other methylphenidate products identified psychiatric adverse events as a possible concern. The review concluded that labeling regarding these events could be improved, and in June this Committee recommended that these adverse events be examined with an eye towards improving the product's labeling regarding these events. And so, accordingly, the Division of Drug Risk Evaluation undertook a review of clinical trial and post-marketing data regarding psychiatric adverse events with these drugs. And this involved, as you've already heard, requesting information from the sponsors, and we asked for both post-marketing and clinical trial data. And we included the sponsors with pending applications, as well as approved applications for this indication.

And we focused on four categories; first of all, psychosis or mania, and we can get into a little bit why we grouped those together, but I think just to state it briefly, as a practical matter because it's very difficult to distinguish within the data whether these were psychotic disorders or manic bipolar-type events. Secondly, suicidal ideation and behavior. Third category was aggression and violent behavior, and then we also had a miscellaneous category where we asked for serious adverse events, but these data won't be discussed today.

For the post-marketing analysis, the time frame was from January 1^st, 2000 forward, and Dr. Gelperin is going to present the post-marketing data in the next talk. And this presentation will focus on the clinical trial findings. So turning now to the clinical trial analysis, the purpose was to characterize the adverse psychiatric events observed in ADHD clinical trials with an emphasis on these three categories, as I mentioned. And to determine the rates of these events by pooling the data, primarily by development program, but also data for all methylphenidate oral products.

And the information requests that went to the sponsors, which I mentioned previously, for the clinical trial data, the sponsors were asked to perform a text string search of their adverse event databases using both preferred terms and investigator verbatim terms, and searching for terms representing events from these categories.

After that, the sponsors were asked to enumerate these events according to exposure, age group, gender, and trial, to provide brief clinical descriptions of the adverse events identified in the search, also to provide descriptions of the clinical trials included in the analysis, and all the sponsors responded, and we were able to review and aggregate the data, and I'll be presenting the findings.

The products we included, this by now is getting to be a familiar list, I think, but basically methylphenidate, we had Concerta, Medidate CD, and Ritalin LA, atomoxetine, Adderall XR for amphetamine, and for d-methylphenidate, Focalin and Focalin XR. For the pending applications we had the methylphenidate transdermal system, and Modafinil. We did receive adult data, although this is described in the written report, which is in the briefing package, but I won't be presenting it in this slide presentation.

Okay. To characterize first the clinical trial data sets, this display is for the double-blind portion of the development programs only, and it shows the number of patients randomized both to drug and placebo. And you can see that in general, there were several hundred subjects in these development programs, and the one notable exception, atomoxetine, had a considerably larger sample size, and followed by the Adderall XR development program. And these were all pediatric patients, by the way.

Now this is a similar look at the data, but here we're looking not at numbers of patients, but at person years of exposure. And as people know, a person year would be one patient taking the drug for one year, or it could be two patients taking a drug for six months, as an aggregated measure. And you can see that here, atomoxetine has by far the largest exposure when measured in this way. Part of the reason is that this includes a long-term relapse prevention trial data. But I think you can also see that for some of the products, the actual exposure times were relatively short, just 20-30 years of aggregated person time in double-blind studies.

Okay. Now let's look at the numbers of events for these three categories, the next three slides I'll be showing you that. And starting with the psychosis or mania events, and this display - this is rates of the events for hundred patient years, and you'll see that the products are displayed here. And this is aggregated, all the oral methylphenidate products were pooled together here. And a couple of things to notice.

First of all, there are very few events, a total of 13 events across all double-blind trials that fit in this category, so that's number one. Number two is, despite that, you'll notice that all 13 events occurred on active treatment, and there were zero on placebo, so that's unlikely to be due to random chance, suggesting some effect of the drugs. And then you see actually the highest individual rate applies to the methylphenidate transdermal system, with four events, and the others, as you see.

Looking now at suicidal events, once again - well, let me add first of all, there are no completed suicides in these trials. And most of these events were ideation rather than actual attempts. You see once again, there were not many such events, that's a total of 17. And then you also see that for a couple of products, there appears to be an imbalance relative to placebo, Modafinil at the highest rate per hundred person years. And then also, atomoxetine an imbalance relative to placebo, and the sponsor has done an analysis of that, which I'll present in a few minutes. And then apart from those, there was one event with Adderall XR, and one on placebo in the methylphenidate study.

And then the third category, aggression - and here we see actually there's far more events, much higher frequency both on placebo and on drug. One point to make is, you can see just looking at the placebo, there's some variability from development program to development program, as far as the placebo rate, suggesting there's some differences in ascertainment, and perhaps population studied. Secondly, looking at this, you'd be hard-pressed to say that there's much evidence that the drug treatment is actually reducing aggressive behaviors in these subjects; although, intuitively one might think that treating ADHD might reduce aggressive behavior in kids, but in these adverse event data, that didn't really seem to be the case. There's only one, in fact, Modafinil, where the rate was numerically lower for active drug. And with methylphenidate transdermal system, actually the rate was considerably higher, although this did not reach the customary level of statistical significance.

Okay. It's of interest to look at the placebo group, because this is a large sample of ADHD patients without medication. And in aggregate, there are almost 4,000 representing over 400 patient years of exposure. As you might expect, these are predominantly males and pre-adolescents. And as I mentioned, the absence of any psychotic or manic events in this group was notable.

For suicidal events, the rate was just under one per hundred of person years, and those of you familiar with the antidepressant pediatric trial analysis, this is a much lower rate than the rates we were seeing in that meta analysis. Of course, this is a different indication, and also, it tends to be a younger population. And aggression events, as you see, were far more frequent than the other categories.

Okay. The next two slides, I want to show you Lilly did their own analysis of two of these categories for atomoxetine. And first, this is the suicidal events. And these findings are actually reflected in the Strattera labeling. And you see that they use two categories of events; one, suicidal events using an FDA definition, and a second one using Lilly's own definition. And this was indications not limited to ADHD, so it's a slightly different set of studies. And also, this did not include the long-term studies that I mentioned, so it's a little different pool of studies from the data I showed you a couple of slides ago. You see, there is this imbalance with six events on atomoxetine, zero on placebo, and then with the other definition, it's seven versus one. And you can see the P-values, this one did reach the customary level of statistical significance.

Interestingly, the adults really not much of a difference between drug and placebo in frequency of these events. So this finding, as I said, is mentioned in the Strattera labeling.

Also, for aggression, similarly, Lilly looked at these events, and you can see there is a slight excess relative to placebo in terms of frequency. And, again, that is now described in the Strattera labeling. And for comparison, this was a methylphenidate active control group, which is somewhat lower.

All right. This is a rather dense slide, but it's just to present the open label data. And this, I guess in a sense, sort of compensates for the relatively short duration of the double-blind studies. You can see that with longer exposure time, we've picked up events in most of the categories for these drugs. And I'll be coming back to the psychotic mania events in a minute.

So what are the limitations of these data? Well, first of all, the small sample sizes, and the short duration of treatment, as I've described. And also, the small number of events made any estimates of the rates relatively unstable. It's possible there's lack of consistency across trials with respect to ascertainment and reporting of these events by investigators. There's a possibility of misclassification of cases. The case definition and search terms, we can't be certain that they had adequate sensitivity and specificity. And I should add that Lilly was the only sponsor that in their own analyses actually adjudicated the event reports.

So just to summarize the findings, for aggression events, it was more frequent with drug than placebo for methylphenidate, patch, Ritalin LA, and atomoxetine. And again, there's little evidence from these data, at least, that the drugs are successful in reducing the rate of aggressive events.

For suicidal events, there is an imbalance for Modafinil and atomoxetine relative to placebo. As I mentioned, the atomoxetine finding is statistically significant, and is part of the labeling now. And for the psychosis and mania events, as I mentioned, we had 13 events, all of them happen to -- in the double-blind trials, all of them happen to curve on active treatment. And in open label treatment, you can see that there was some frequency with all of the products, and the highest percent was the methylphenidate transdermal patch.

And this is just to provide some examples of some of the more clinically significant events from this category. There was one 8-year old treated with Modafinil, and a 9-year old treated with atomoxetine. Both of those children were hospitalized. Two other cases with Adderall XR, the methylphenidate patch were not hospitalized, but the drug was discontinued by the investigators. And in this case, the child on Modafinil had a previous history of psychotic symptoms which had not been elicited at the time of enrollment into the study. So that's just to illustrate the types of events that were in these categories.

Some more comments on the trials, many of these trials were of short duration, and also in terms of the patient population, many enrolled subjects previously known to respond to drugs in the class so that these factors limit the utility of these trials for determining the drug product safety profiles. In particular, if children known to respond adversely to the drugs in a class are excluded from the trials, then that, perhaps, maybe not be generalizable to the population for this indication.

And then just to close, a few brief comments on some literature reports. I have these two slides. This first one is case reports, and you can see there is one case series of six. These were adults, but six patients admitted to the same psychiatric hospital in a three-month span with psychosis thought to be related to Dextroamphetamine, and that caught the author's attention, so they wrote it up as a case series. There is description of two children who developed what is termed toxic hallucinosis on methylphenidate, and similarly, another case series of three children with visual and tactile hallucinations on methylphenidate. And a report of a 13-year old female who was described as having typical amphetamine psychosis with Adderall treatment. And then some publications that had sort of more systematic data collection. Cherland and co-authors did a chart review of 98 child outpatient's diagnosed with ADHD, and treated with stimulants, and they found that 6 percent developed psychotic symptoms. That's a much higher rate than what was seen in the clinical trials I just described. And Dr. Gelperin is going to have some more comments on this.

Similarly, for atomoxetine, a case series of 153 sequential pediatric outpatients, the incidence of mania was 6.5 percent. Again, higher than what was seen in the clinical trial population. And again, Dr. Gelperin is going to mention that. And then there was a placebo-controlled randomized study of amphetamine with divalproex sodium in patients who had pediatric ADHD and bipolar disorder. It was a very small study, but it was interesting to note that none of the 30 subjects had worsening of mania when they were receiving concomitant Valpro-8 plus the amphetamine.

So I'll stop there, and I want to acknowledge, first of all, the sponsors who provided their clinical trial data, and also the people who assisted at the FDA. And with that, it's my pleasure to bring back again Dr. Kate Gelperin, who's going to talk about the post-marketing data for psychiatric adverse events.

DR. GELPERIN: Thanks, Andy. Today, I'm going to tell you about the results of an analysis of psychiatric adverse effects during drug therapy of ADHD. I'm going to talk about the methods that we used for the case review and the analysis. I'm going to talk about the MedWatch reports that reviewed the three main categories that Dr. Mosholder had discussed. I'll talk a little bit about clinical implications, and what might be some conclusions.

The post-marketing safety information was requested from the manufacturers of these drugs, and they performed an analysis of each case that they identified in the searches that we described for spontaneous or literature reports that they had received since January of 2000. There were four broad categories of psychiatric adverse events. This discussion today is not going to include the miscellaneous category. And a high level analysis of patient characteristics and potential risk factors for psychiatric adverse events was completed. By "high level", I mean that the cases received the initial hands-on review, and the characteristics were entered into Excel spreadsheets. These data were summarized using SAS and are analyzed as we'll discuss today. However, additional in-depth analyses could also be undertaken for, for instance, more clinical review of individual cases.

At the same time, the FDA Office of Drug Safety undertook a search of the AERS safety database for the same time period, and the identified MedWatch cases were assessed by a team of reviewers in the Division of Drug Risk Evaluation. The reports were classified into the categories, and each report could qualify for more than one category. For instance, a child could exhibit aggressive behavior and have hallucinations, so that case would count in two of the categories.

The hands-on review of the MedWatch reports identified duplicates, and reports which were considered to be of very poor quality, or were highly unlikely to be related to the drug of interest, and they were excluded from the analysis. The criteria that we identified up-front to look at the characteristics of these reports, since they were quite large in number, were adapted from Neuron Ho and a publication that was originally intended to describe a rating system for individual cases, and has since then become really incorporated into post-marketing drug risk assessment as some of the cornerstones that Dr. Iyasu also described to you to help to sort through the questions about post-marketing data, and really make the most of the individual case reports that come to us spontaneously. So, for instance, we looked at are there published case reports that might be consistent with a causal association? What's the temporal association? Did the events improve or resolve when the drug was stopped? Did the events recur when the drug was re-administered? That's the positive re-challenge, and that's one of the hallmarks for post-marketing causality assessment of spontaneous reports.

The cases were also assessed for alternative factors that could cause or contribute to the adverse event, such as concomitant medications, drug abuse, or a pre-existing condition with similar signs or symptoms. We also considered whether the case was confirmed by a health professional or originally reported by a health professional. We examined various demographic factors and other potential risk factors that our colleagues in the Division of Psychiatric Products were interested in, such as were there seizure disorders in these patients, was there a psychiatric history other than ADHD? Was there a family history of serious psychiatric illness? And in the time constraints today, I'm really going to just be able to give you just a very high level overview.

These are the search terms that we were interested in for the category of psychosis or mania. Hallucinations of any type we were interested in, events that were coded as psychotic disorder or acute psychosis, paranoia, and mania.

Now I want to orient you to this slide, because you'll see this format again. Okay. So the drug names, you've seen these numerous times today, and I am going to use generic names. I'm going to use the name amphetamine to refer to all branded or generic amphetamine or dextroamphetamine products, either short-acting or long-acting, that would include Adderall. It would also include Dexedrine. Atomoxetine is the brand name Strattera. I'll be referring to it as atomoxetine. Methylphenidate, these numbers include all brands, all formulations currently marketed, such as Concerta, Ritalin, both short-acting and long-acting.

This drug, Modafinil, the brand name is Provigil, is in a gray box because it is not currently approved for ADHD. It is, however, currently approved for narcolepsy. And so the sponsor was asked to send us the information about the narcolepsy patients, and we found that there is some small amount of off-label use in ADHD, and so these numbers are included here for completeness, but I'm really going to focus in the time that we have on the currently approved drugs.

Also, I am presenting to you the results of the FDA AERS analysis. You did receive in your background package a 90-page document that goes into excruciating detail of the sponsor's analysis and the FDA analysis. And I'm happy to say that on all of the points worth mentioning, that were pretty much in agreement, so today I'm going to tell you about the numbers from the FDA analysis.

These are for the category psychosis or mania. These are the non-excluded cases, which means we thought they were of sufficient quality to be included in our analysis today. I will also point out to you that there's an N, which is the count of cases, and we are not attempting to use reporting rates or to otherwise try to assess the frequency of the occurrence of these events, either relative to other drugs that are used in this condition, or relative to expected rates for a variety of reasons, we did not feel that would be appropriate. So even though these numbers look different, we are not making a conclusion based on that at this time.

Patient age, you will see that the majority of the reports of psychosis or mania for the drugs currently approved for ADHD were in children and adolescents, predominantly male. Fewer than a third typically had another psychiatric history, and we were surprised, there were very few with drug abuse on board, and very few with overdose, and very few with seizure disorder. NR means that that information was not reported. That could mean that it didn't happen, it could mean that it did happen but the reporter forgot to tell us.

Sort of the same setup with the drugs, but some different attributes that may help us to understand whether we think there's a drug relationship here. We wanted to look at how many of these reports are case reports from published medical literature, and we found that yes, there are some. How many get better when the drug is stopped? Well, maybe a third to a half. Now in many of these cases, they may have ultimately gotten better, but at the time the case was reported, they hadn't gotten better yet, so this number should not be taken to mean that these are the only ones who got better when the drug was stopped.

Positive re-challenge is the gold standard for post-marketing causality assessment. That means the drug was stopped, the event went away. The drug was given again, and the event came back. And we did have cases of that sort. Many of the cases did not have concomitant medications on board which would be potential confounders, and it was also striking that almost 90 percent of cases across the board did not have a prior history of a similar psychotic event. Medical confirmation was obtained in half to three-quarters of these cases.

I'll just comment that the Modafinil patients were typically older, they were adults, typically treated for narcolepsy and there was some occurrence of psychosis related events.

I'm just going to give you a couple of examples of cases to give you a sense of what are we talking about. This is an amphetamine case from published literature, Journal of the American Board of Family Practice in 2002. A 12-year old female who developed hallucinations after five weeks of therapy with 10 milligrams a day of amphetamine for the inattentive type of ADD. She was on no concomitant medications. She didn't have much in the way of family history or medical history. The amphetamine was stopped and Clonazepam was started PRN for agitation. However, she got much worse, and four days later she really was having visual hallucinations, command auditory hallucinations, and tactile hallucinations of bugs crawling under her skin. She also displayed waxy flexibility, and she was admitted to the hospital. She was kept free of medications and she returned to her baseline within seven days. She was fully evaluated and a tox screen was negative, so the authors concluded that this was an amphetamine reaction, even though the dose was not excessive, and she was discharged home on no medications.

This is a report from the sponsor describing a 6-year old male who was started on 100 milligrams a day of Modafinil for ADHD who experienced visual hallucinations after a single dose, and they did resolve when the drug was stopped. There were no concomitant medications.

This is a report from the sponsor which describes a 7-year old female who received 18 milligrams a day of atomoxetine for ADHD. Within hours of taking the first dose, the patient started talking non-stop and stated that she was happy. The next morning the child was still elated, two hours after taking her second dose she started running, stopped suddenly and fell to the ground. She stated she had run into a wall, but there was no wall there. The atomoxetine was stopped, and this report did not include the outcome of the events.

This report is a methylphenidate case from published literature from The Journal of Neurology from 2004. A 12-year old boy with cerebral palsy, low normal intelligence and a combined sub-type of ADHD was treated with 10 milligrams a day of methylphenidate, and his attention improved. However, one morning he was observed crawling on the floor complaining that roaches were surrounding him. This phenomenon appeared two hours after his dose, continued for two hours, and then disappeared with no intervention. Methylphenidate was stopped; however, his school performance deteriorated, and so the methylphenidate was resumed. However, the hallucinations immediately recurred and were stopped, and three years of follow-up have been uneventful.

So the question is, how often do such things occur at usual doses? So Dr. Mosholder showed you, for instance, the proportion of patients in open label long-term extensions from the clinical programs, and all of those rates were less than 1 percent with the exception of the methylphenidate patch. Interestingly, this chart review, which is actually the only similar information I could find in published literature for stimulants, describes the experience at an out-patient clinic in Canada from the years 1989 to 1995.

In that five-year period, of 192 children who were diagnosed with ADHD, 98 were started on stimulants, most received methylphenidate. Of those, six developed psychotic symptoms, and so giving us a frequency of around 6 percent. One hypothesis is that since only half of the children with the diagnosis were treated, perhaps these children are somehow more severely affected, so maybe this is the upper bound. So maybe, we might think that the rate is somewhere between under 1 percent, up to 5 percent, or 6 percent.

This is for atomoxetine, a publication in The Journal of Pediatrics from the year 2004 of pooled data from out-patient settings in Colorado and Minnesota. The author is a pediatric psychiatrist. He described a total of 153 sequential patients treated with atomoxetine, of whom a third developed unwanted psychiatric symptoms. Of that third, however, 80 percent did have a past history of mood symptoms, so again, we may be talking about a more severely affected population, or perhaps some additional co-morbidities. Of these, ten children developed symptoms severe enough to be considered mania, three requiring hospital admission, and three incarcerated in juvenile detention centers.

So what are the findings? We found no risk factors that could account for the majority of reports. For instance, drug abuse was not a common feature, and the vast majority of these reports did not have a history of a similar psychosis-related condition. There were positive re-challenge cases identified, and many cases had a positive de-challenge. This may not be a rare occurrence based on the published case series, and to some extent, even from the clinical trial experience.

A large proportion of these cases do involve young children, and in our review of the narrative, it was striking that often the young children were describing various insects, bugs, snakes, and worms with a combination of visual and tactile hallucinations, which we haven't seen described elsewhere.

So the labeling, I'm including here just for your reference. Dr. Andreason told you a bit about this earlier, and you will be asked to look at this and advise us later. But to summarize, the current approved labeling for drugs with the ADHD indication does not clearly address the risk of drug-induced signs or symptoms of psychosis or mania, such as hallucinations in patients without identifiable risk factors, or occurring at usual doses. So this is a setting where there's no previous psychosis, and there's no overdose. Current labeling does not address importance of stopping drug therapy in patients who develop signs or symptoms of psychosis or mania. And we will ask for your advice later today.

Now on to aggression or violent behavior, these were the search terms that we requested, aggression, anger, hostility, homicidal ideation, and even murder or imprisonment. This is the similar format that you've seen. And again, you can see that the predominance of the reports is in children and adolescents, predominantly male. Psychiatric history in a quarter to nearly a half, although some of these were fairly soft conditions. Drug abuse, interestingly, in very few, overdose in relatively few, and very few with seizure disorders.

Published literature for methylphenidate and atomoxetine during this five-year period, I wouldn't be surprised if we expanded our search dates, maybe we would find something for this. Positive de-challenge in some cases, positive re-challenge cases were identified, many with no concomitant medications, although for some reason, the atomoxetine group seemed to have more concomitant medications on board.

We were a little surprised, given ADHD diagnosis, that the striking majority of cases reported that they didn't have a prior history of aggression or violent behavior, so we're not sure what to make of that. And many of these cases did have health professional confirmation.

So the findings, most cases were classified as non-serious, although about 20 percent were considered life-threatening or required hospital admission. There were some juvenile incarceration cases. Most of the reports did involve children and adolescents, and we did not identify any specific risk factors that could account for most cases, such as drug abuse. And there were some positive re-challenge cases.

With regard to the labeling, the amphetamine and the methylphenidate products don't currently have labeling about aggression at the usual doses. However, the Strattera label does have that, as was discussed previously.

On to suicidality, these were the search terms that we requested, suicidal, depression, gunshot wound, intentional self-injury, non-accidental overdose, self-injurious behavior, self-injurious ideations, self-mutilization, suicidal ideation, suicide attempt, and completed suicide. We did have some completed suicides in the post-marketing data unlike in clinical trial data where there were no completed suicides.

Similarly, for the approved drugs, we did find that a large proportion of these cases were in children and adolescents, male predominance. For these suicidal ideation and suicidal behavior cases, maybe there were a few more with some other psychiatric history reported. Drug abuse was, perhaps, a little bit higher than on the previous slides, but still not in a majority. Overdose in many of these was actually the method of the suicide attempt or behavior, very few were seizure disorder. Published medical literature, yes, for atomoxetine and methylphenidate - again, if we opened up the dates and looked further, maybe we would find something more for that.

Positive de-challenge in a little more than a third of the cases, and there were some positive re-challenge cases. The majority had no concomitant medications, and the majority again did not have a prior history of similar events. Many cases were from health professionals.

So the current labeling for the amphetamine and methylphenidate products does not include any information about suicidality. We felt that a possible causal association between stimulant therapy and suicidality couldn't be ruled out on the basis of our analysis. With regard to atomoxetine, we felt that the results of the post-marketing review were consistent with the association that is already clearly described in the labeling.

So to summarize, suicidality has been identified as a potential safety issue for atomoxetine and that information is conveyed in the labeling. A causal association between other drug therapies of ADHD and suicidality could not be ruled out on the basis of our analysis. We recommend further evaluation, such as a clinical expert case review, for some of these cases that were identified regarding a possible co-occurrence of undesired psychiatric effects that perhaps could contribute to suicidal ideation or behaviors in vulnerable patients.

With regard to aggression or violent behavior, numerous post-marketing reports have been received, most in children and adolescents with a male predominance. We did not identify any specific risk factors which could account for the majority of the cases that we identified, and these data suggest that some cases, aggression or violent behavior may be drug-induced. The Committee will be asked to discuss any labeling implications later today.

With regard to psychosis or mania, we found that signs and symptoms of psychosis or mania, particularly hallucinations, can occur in some patients with no identifiable risk factors at usual doses of any of the drugs currently approved to treat ADHD. Based on published case series rates, and to some extent the clinical trial data, this may not be a rare occurrence. No risk factors were identified, which could account for the majority of reports of psychosis-related events. For instance, drug abuse was reported in two cases.

The predominance in young children of hallucinations, both visual and tactile involving insects, snakes and worms, we think deserves further evaluation. The Committee will be asked to discuss any labeling implications of these findings later today.

We'd like to thank the manufacturers of the drugs discussed today for providing timely and comprehensive safety data for these analyses, and our colleagues in the Division of Psychiatric Products.

DR. NELSON: Thank you. And let me commend all of our speakers for doing what I was highly doubtful could in fact occur. Everyone stayed on time, amazing.

Now the second point is, we're scheduled for lunch at 12:25. We were scheduled for 10 minutes of clarification questions. We don't necessarily need to take 25 unless we want to. I'll point out at least for the Committee, we have box lunches coming in. The only reason to give a longer lunch break is the people in the audience, I suspect, do not and might want to try to get back at 1:00 on time, so we'll see how much time you need to take for that. And I will also, before opening up for clarification questions to our speakers, is to point out to the Committee that everybody that has presented, it's my understanding, will be here during our discussion this afternoon, so there should be no need to get their insight into the questions themselves. All we need to do is focus on informational aspects for the moment, because they'll still be available to us this afternoon. So with that, just open for clarification questions about any of the information that we've been presented this morning. I'll start keeping a list. John, you're first, then Judith, and then --

DR. MOORE: A question for Susan McCune and Dr. Gelperin. One year post-exclusivity adverse event for you on Adderall, well, the Adderall XR and Adderall both, you mentioned seven cases total, I think. And then Dr. Gelperin, in what I think is the same data, mentioned only four cases of sudden death. Could you two kind of reconcile that for us?

DR. GELPERIN: Yes. I know Dr. McCune tried in her beautiful slide with the Xs and the Os to explain this, but it's minutia of post-marketing report counting. The issue is that two additional of the cases that were presented by Dr. McCune were previously reported in the prior time period, but because of the conventions of searching for the BPCA, the safety evaluator who did those searches did not exclude those cases from her analysis. However, I felt that it was not reasonable to include them in the reporting rates when they were already included in the reporting rates for the other time period. And then the third case, as we mentioned, is the one where the amphetamine was stopped two months prior to the death, and the child was actually on Strattera and Buproprion at the time of death. That could have been -- you could go either way with that, I agree, but we decided not to include it in the reporting rate.

DR. NELSON: Judith.

DR. O'FALLON: In all the data that was presented about the efficacy, which I found very helpful, there was one that was mentioned that the efficacy rating was done on the basis of the clinician. And that one kind of blew my mind, so I wondered how is it that the clinician can assess efficacy? I understand what the parents are doing, I understand what the teachers are doing. I don't understand quite what the clinicians are doing.

DR. NELSON: Ben.

DR. VITIELLO: The clinician basically collects information from the patient, and when it is available from family members, and so makes a determination based on that information.

DR. NELSON: But I assume that they were blinded to treatment assignment in those studies. Right?

DR. VITIELLO: Yes, that goes without saying. It was a double-blind assessment.

DR. NELSON: Dan.

DR. PINE: I have three quick questions for Dr. Mosholder. First of all, for the adverse events data in the randomized controlled trials, I didn't see statistics for the mania in the stimulants versus placebo, the 13 versus zero. Could you give either an odds ratio, risk ratio, lower bound confidence interval, something like that?

DR. MOSHOLDER: Yes. I did that sort of back of the envelope on my own. I didn't put it in the report, because it's a little unsatisfactory to do that sort of crude pooling without a formal meta analysis. I think it was out to three decimal places, but I didn't make note of that. To really do a more precise meta analysis and I'll defer to the people who are more expert in statistics, I thought that the data here were too sparse, and you'd be left with a lot of zero cells, and you'd have to make corrections for that, so I just sort of didn't go there. But it would have been -- if you just crudely pool everything together, I think the P-value is out to three decimal places, but I don't have that, unfortunately, on me at the moment.

DR. PINE: And what about a point estimate on an odds ratio or risk, did you -- that's all right.

DR. MOSHOLDER: Yes. Well, that would be undefined because of the zero for placebo. You could put a confidence limit on the rate difference, but I don't have that.

DR. PINE: So the second question was about the data on aggression, and I wondered the degree to which you had available ratings that were obtained during the study of aggressive behavior, because obviously, most of the studies would obtain those. And it would be important and interesting to see if those data look similar or not to spontaneous reports of aggression; so do you have those data?

DR. MOSHOLDER: I did not for this analysis, although I will say that in Lilly's analysis of aggression, they did look at some of the efficacy rating scales which included items relevant to aggression. And I don't think there was much of a finding either way, but it is my impression that some of the standard efficacy ratings used in these trials has such items, and maybe not all of them did, but you're right - that would be another way to -- and to be fair, these trials were not designed to look at aggression as an end-point in itself. It was simply observing aggression counted as an adverse event, is the way I looked at it.

DR. PINE: And then the last question concerns an opportunity to either query the data set or query the companies for some inconsistencies in some of the case reports of hallucinations or psychosis. So, for example, I notice that a couple of them that you showed were listed as psychosis, but then were listed as not serious, which seemed a little curious. Was there any opportunity to try to clarify that discrepancy, or figure out why that appeared like that?

DR. MOSHOLDER: Certainly, that could be done. I didn't have all the information or the time to delve into all of the cases, but I guess you're raising the question of how could you have a psychotic event and not consider it serious? Well, that's up to the investigator and the sponsor to see if it meets the regulatory definition for serious. I think the other point is that the distinction - a lot of the events clearly could have fit into more than one category, and that someone having command hallucinations to do self-harm, that could be counted either as a psychotic event or a suicidal event, and so in a certain sense, some of these distinctions were a little arbitrary.

DR. NELSON: Bob Ward.

DR. WARD: I had a couple of questions for Dr. Gelperin. In the suicidality slide, if I recall that table correctly, about 35 percent had a prior history of suicidality, either ideation or attempt, and roughly 45-50 percent were not reported. So if you took away the absence of data, the majority would have a prior history.

DR. GELPERIN: Okay, just looking. Are we talking about slide --

DR. WARD: I don't know what your slide numbers are.

DR. GELPERIN: Oh, okay.

DR. WARD: Put your slides up, maybe we can tell.

DR. GELPERIN: Can we put slide 28 back up? Okay.

DR. WARD: It was a table, it was in that same format you had been presenting for several of the data tables.

DR. GELPERIN: Right.

DR. WARD: Oh, I think I saw it. It was the one that had - yes, okay, right there. Back one. There. When I scanned through there, it looked like other psychiatric history yes, for almost all, unless there was an absence of data, not reported. Correct?

DR. GELPERIN: This was -- so each individual case was assessed either by the sponsor or by the FDA review panel. And then a yes or a no, or a not reported could be the answer. So this tabulates the answers, and typically in a post-marketing spontaneous report, it's like if somebody were telling you a story. The things they remember to tell you, you hope they're remembering the important things, but they might forget some important things, or they might stop talking to you before they tell you the important things. So that's why not reported is a big category.

DR. WARD: It is a big category.

DR. GELPERIN: Sometimes what that means is that if it were really important, you might think that someone might have mentioned it. So, for instance, this is the FDA analysis you're looking at. If it wasn't stated, we said it was not reported. Some of the sponsors actually in their analysis kind of made the leap that if some -- that they wouldn't forget to say oh, by the way, this person has schizophrenia, and now they're suicidal.

DR. WARD: What I would, I guess, prefer to see you address would be if you take away those where we don't have data, it's not reported, it would appear that the predominant number of these cases had some prior psychiatric history reported.

DR. GELPERIN: So 45 percent --

DR. WARD: If you just took them out of the analysis all together if they didn't have any report, at least that's the way I would have to deal with it in the absence of data.

DR. GELPERIN: Okay. So, for instance, let's look at atomoxetine. So roughly a third stated there was another psychiatric history. I can tell you that those were typically things like oppositional defiant disorder, obsessive compulsive disorder, but for some reason, the sponsor also decided in some analyses to include things that were sort of softer. So it's not -- that shouldn't be taken to mean that these were DSM-4 criteria, psychiatric diagnoses. Some of them were pretty soft things that just weren't part of the ADHD spectrum. But I actually thought your question was really on the next slide, which was how many reported a previous history of a similar event.

DR. WARD: No. Actually, it pertained to that previous slide.

DR. GELPERIN: Oh, the other psychiatric history.

DR. WARD: Yes. The majority, for whom we had data mentioned or not mentioned, that there was a large percentage that had a prior psychiatric history, and that we know that ADHD may be a component of multiple behavioral and psychiatric spectrum disorders.

DR. GELPERIN: Actually, for the post-marketing reports that would not be the correct conclusion, because the only reports in this category - so there's another one that would have said no. There would have been a no, and that would be a report where the reporter stated specifically there was no other psychiatric history. So not reported means they didn't say that, no means they said there was no other psychiatric history, and yes means they provided some other psychiatric history. So you could not conclude that the majority probably had other psychiatric history.

DR. WARD: Then the column we're discussing atomoxetine, the no would be 1 percent?

DR. GELPERIN: That's right.

DR. TEMPLE: Kate, that's the point he's making, that where you have information it's almost always yes. In other words, for the first one, you have information on 90 percent of patients, and apparently they said no in 10 percent.

DR. WARD: Right.

DR. TEMPLE: So, I'm sorry, you have information in 55 percent, 45 percent of the 55 percent is yes, 45 percent didn't say anything. I think that's the point you're making.

DR. WARD: Yes. Thanks, Bob. Yes.

DR. TEMPLE: Where you have information, it's usually yes. Now then you have to figure out what the no reports mean. Your thought is that it's probably no, otherwise they would have reported it. But that may or may not be the case.

DR. NELSON: So I've got Dr. Daum, and then Dr. Armenteros. Robert.

DR. DAUM: Thank you. So my question is also for Dr. Gelperin. I was hoping for some clarification with the category of aggression or violent behavior. I'm sorry for asking that question back to you, but that's where I'm proceeding. It seems to me that there's quite a wide range of behaviors there, from aggression, however that was defined, all the way to murder. And that those behaviors may or may not have different weights in terms of their social acceptability, to say the least. So when you get to the slide where there's challenge and de-challenge data, can you give us more information? I mean, would someone actually commit a murder and then was re-challenged? I don't mean it facetiously, but --

DR. GELPERIN: No. That's a great question and that was really what I was alluding to when I said our time constraints were such that what you were hearing about is just such a high level, but there's obviously layers and layers in these data. I can actually answer your question specifically with regard to the pediatric re-challenge cases for atomoxetine. So there were 12 such cases in which a positive re-challenge occurred. I don't have a slide for this, but I can just tell you that they were all classified as non-serious. They were -- the Medra PTs were aggressive for four, anger for seven, and hostility for one. It was considered to be an exacerbation of a pre-existing behavior in one case, and it was not in 11 cases. There was no drug abuse in 12 cases, and the age ranged from 4-14 years, with a mean of 9.6 years. Six were males, four were females. There were no concomitant medications in four of the cases, and the daily dose of atomoxetine ranged from 10 milligrams to 60 milligrams. Five of these were healthcare professional reports, and seven were consumer reports.

DR. DAUM: Can I try to probe a little more for what I'm after. Can you tell us a little more what hostility means in this context, and what anger means, because I'm sure no one in this room would ever be accused of being hostile or angry, but why were these reports filed for hostile or angry behavior? It must have been pretty impressive.

DR. GELPERIN: Actually, it's quite striking the number of reports that have been filed for atomoxetine in this category. The vast majority are non-serious reports, but many of them are from healthcare professionals. And we actually don't know why this is happening, but for somebody to pick up the phone or to tell the sales rep, or to -- it means that somebody in the United States decided that the drug company and FDA should know that their child started hitting somebody and didn't do it before.

Now let me give you another example of a positive re-challenge case that is a more serious case, and shows the overlap, and perhaps gives a sense of how the co-morbidities may occur. This was a consumer report, but it's the mother of a 9-year old boy who reported that her son started hitting people and cutting himself, and he was hearing voices after a few weeks of treatment with Adderall XR for ADHD. The mother decided on her own to stop the medication and the events resolved. However, the physician advised that she should restart it when school started because the boy needed it. The dose was then increased to 15 milligrams daily one month later, and the events returned, so that's a positive re-challenge.

Seraquel was added to the child's treatment because of the auditory hallucinations, however, the child worsened and was admitted to a psychiatric hospital. So I think that gives you a sense of the vast majority of these cases are non-serious. Also, for post-marketing spontaneous reports, the quality of the information is not like a published case report. We are basically listening to a mother who's telling us something. Did she get it right? I don't know, but it's my job to listen to what she's telling us.

DR. NELSON: Jorge.

DR. ARMENTEROS: Don't go too far. Actually, pertaining to the same topic. I believe the ascertainment and the search terms used included, as you mentioned, a wide range, some of which are pretty serious. But perhaps one, and I wonder if you had the opportunity of looking at it, is more simple and maybe mediating some of this, which is agitation. That could actually be a component behind the most common episode that you described. Any thoughts about the possibility of looking at that?

DR. GELPERIN: The most common component of what?

DR. ARMENTEROS: What I was saying is that the possibility of looking at agitation in the process of evaluating your aggressive repertoire.

DR. GELPERIN: I think that's an excellent point. And really, I would almost make a plea that additional analyses should be done, because if you look at our time frame, the Committee asked us to do this analysis only a little more than six months ago, and so it takes a long time to get the sponsor, the data searches, the Excel spreadsheets. I mean, a thousand cases had to be reviewed by groups of people who are expert in this, but it was really - I would call this a high level analysis, where we selected using the Neuron-Ho criteria, we selected key points to help us assess the quality of the information, but I think that the clinical analysis, and in-depth analysis of these issues - and, in fact, I really wonder whether something like the multi-modal, or some of the trial data could be sort of plumbed a little bit further to really help us out with understanding the rates and maybe co-morbidities.

DR. NELSON: Tom.

DR. NEWMAN: A question for Dr. Mosholder. I'm wondering what the miscellaneous serious adverse psychiatric effects were, and if you know the distribution between the drug and placebo groups, what those were?

DR. MOSHOLDER: Yes. Unfortunately, this will be a very brief answer, because I don't have those data really analyzed. I think some of the serious events were actually more in the behavioral realm, but were coded, for example, as personality disorder or something like that. And I have to say, too, one complication was that some of the sponsors misunderstood that we only wanted serious outcome, regulatory serious outcome so that in some cases we got almost all psychiatric adverse events fitting any of the miscellaneous criteria, regardless of whether they were serious, so that was a little overwhelming to try to sort through, so I don't have those data.

DR. NEWMAN: You don't know whether there were more with drug or placebo.

DR. MOSHOLDER: No. And actually, the miscellaneous category really spanned quite a range from tics to insomnia, so I wouldn't even want to -- I would want to parse that out before I started comparing frequencies.

DR. NEWMAN: One other quick question, the suicidal events, do you have a feel for how many of them were ideation versus attempts or behaviors?

DR. MOSHOLDER: Most of them were -- I can't give you the exact number, but most were ideation, and fewer attempts.

DR. DIAZ: In the adverse event reports for Adderall XR, there are two cases reported of disability. Can you expand, describe what those were, please?

DR. McCUNE: You're talking about the two cardiac events with cardiac disability?

DR. DIAZ: Yes.

DR. McCUNE: Yes. Those were actually two patients that had underlying cardiac disease that actually had adverse events associated with cardiac surgeries.

DR. NELSON: Ben, I'll give you the last question before lunch.

DR. VITIELLO: More of a question, was just a comment. Since the issue of hostility and aggression has come up, clinical trials indicate that there is a treatment effect on that, but in the treatment decreasing, hostility and aggression, that actually is statistically and clinically significant effect, which indicates that these events of aggression and hostility that has been reported are really idiosyncratic episodes that occur against the mean, which is actually the opposite effect, that these drugs at the group level, they decrease hostility and aggression. This is sort of an important thing to keep in mind.

DR. NELSON: Thank you. We're scheduled for lunch. One o'clock is when the open public hearing occurs, and just so that everyone is aware, we have I think 42 people who want to speak to us, so we will start at 1:00.

DR. JOHANNESSEN: I'd also just like to add that if you are pre-registered as an open public hearing speaker and you haven't checked in at the desk to get your number, you need to do that. The other thing is that we're bringing lunch just for the Committee, so we'd certainly appreciate it if you'd leave them alone during lunch and let them have a little peace and quiet. I think the press office people are here, and they're going to be arranging things for after the meeting. Thank you.

(Whereupon, the proceedings went off the record at 12:25 p.m. and went back on the record at 1:01 p.m.)

DR. NELSON: So if we could start to begin to get organized and I'll start to describe our process.

(Pause.)

So as I mentioned, this is going to be the start of our open public hearing. I know I have to read something before, right? It's right here.

Now I'm going to describe the process briefly, and then I'll read the statement that I need to read before the open public hearing. As you can tell by the slide up front, we have 42 people who are going to speak to us this afternoon. If you divide that into the amount of time, it's basically three minutes each. We'll have people assisting the order and everyone, I gather, has a number.

There's also a timer up there which will be green for two and a half minutes and then turn yellow for 30 seconds. At three minutes, the microphone is off. Now one could say is that fair? Well, from a distributive justice point of view with 42 people, yes. We'd like to give you as much opportunity to get your points across within your three minutes.

So let me read the statement before we open our public hearing and then Jan will take over in terms of calling up people and we'll be operating whatever slides people have from up here.

Both the Food and Drug Administration and the public believe in a transparent process for information gathering decision making. To ensure such transparency at the open public hearing session of the Advisory Committee Meeting, the FDA believes that it is important to understand the context of an individual's presentation. For this reason, the FDA encourages you, the open public hearing speaker, at the beginning of your written or oral statement, to advise the Committee of any financial relationship that you may have with the sponsor, its product or if known, its direct competitors. For example, this financial information may include the sponsor's payment of your travel, lodging or other expenses in connection with your attendance at the meeting. Likewise, FDA encourages you at the beginning of your statement to advise the Committee if you do not have any such financial relationships. However, if you choose not to address this issue of financial relationships at the beginning of your statement, it will not preclude you from speaking.

DR. JOHANNESSEN: So we're ready for speaker number one.

DR. DILLER: Let me know when I can start. I have no financial considerations to take into account here. I'm a behavioral developmental pediatrician. My name is Lawrence Diller who has prescribed Ritalin and other stimulants for nearly 30 years. I've never been against medication. Yet, I applaud the courage of the previous Advisory Committee in broadening the discussion of the use of stimulants in our country beyond the cardiovascular consequences.

The majority's recommendation for a black box warning was also meant to express their concern about the 1 in 10 11-year-old boys taking drugs like these in this country, and that many children have a serious psychiatric disorder requiring medication.

Indeed, in our country, Ritalin and Adderall have become lifestyle drugs, performance enhancers, rather than solely a medical treatment. To that end, even though the risk of sudden death are minimal to existential, these drugs have to be extremely safe if we are giving them to children who are generally very normal and healthy. And indeed, the historical record of 70 years suggests that they are.

Nevertheless, their over-use highlights an ethical, not medical dilemma, because stimulants are not the moral equivalent to helping parents and schools address the needs of these children. But I must alert you to the real and present danger of prescription stimulants. For years, I and a few others have been deeply worried about the risks of prescription stimulant abuse especially in teens and young adults. Now in the February issue of Drug and Alcohol Dependence, we have the first hard data on the misuse and abuse of prescription stimulants. Based on a 2002 Government survey, this data has been vetted by the Drug Enforcement Administration.

Twenty-one million people over the age of 12 have illegally used prescription stimulants at least once. Three million have misused prescription stimulants alone. Of those, approximately 1 in 10, or 75,000, between the ages of 12 and 25, have gone on to meet DSM-IV criteria for addiction or drug abuse.

The use of prescription stimulants are growing the fastest in the adult population, so these numbers are probably larger in 2006. Seventy-five thousand prescription drug addicts dwarf the current cardiovascular casualties. These numbers represent the fourth episode in our history of a doctor prescribed stimulant abuse epidemic. And if our previous experiences are any guide, then several years from now you are likely to look back at this time with regret, chagrin and anger.

Those who do not study history are condemned to repeat it and you may remember the graffiti on the walls of the San Francisco Haight-Ashbury in the late 1960s, "speed kills." We must immediately address this new Ritalin abuse epidemic in our teens and young adults, with much tighter controls on prescription and diversion to this population. At the same time, we need a rational policy for the use of these drugs with a much smaller group of pre-teens who can and do benefit from Ritalin safely.

Thank you.

DR. JOHANNESSEN: Thank you. Is speaker two here? This is Douglas Tynan.

DR. TYNAN: Right. Slides, please. Thank you. Good afternoon, my name is Douglas Tynan and I have no other sponsors. I'm the Director of the ADHD Program at the Nemours Pediatric Clinics in Delaware. I am a child clinical psychologist. I also chair the current task force on stimulant medicine use established by the Delaware House of Representatives.

There have been serious concerns in the First State regarding prescription of stimulant medicines for ADHD. According to the Federal DEA 2002 Report, Delaware has the second highest use of amphetamines and fifth highest use of methylphenidate. Last year's national phone survey of children's health indicates 7 percent and 3.5 percent of girls in Delaware were medicated for ADHD.

Next. Because of the trend of higher use of these medicines, the Delaware House of Representatives established a task force to study the patterns of treatment of ADHD in Delaware of school-age children. We've included parents, educators, public health, law enforcement, health and mental health professionals who have reviewed research and clinical literature, interviewed families, school personnel and national experts.

We have determined that simply curtailing discussion of ADHD medicines in schools is not a practical solution, that we need to do more. Our task force has recommended so far prescription monitoring in the State of Delaware; the Department of Education to establish specific guidelines on how to refer a student needing an evaluation; the Department of Ed. to continue its goal to educate all school personnel and training teachers in behavior modifications for the positive behavior support program from the U.S. Department of Ed.; health insurers provide full benefits of evidence-based behavioral therapies and psychological diagnostic testing as part of mental health benefit packages; and the schools' health and mental health providers to work collaboratively to address any communication barriers that interfere with treatment.

Nemours is one of the nation's largest pediatric subspecialty groups. The practice is moving forward to address this issue. Nemours has developed -- next slide -- as part of the electronic medical record an electronic prompt for pediatricians to use the full AAP criteria, including history, documentation of problems, web-based questionnaires and documentation of impairment before making a diagnosis. The AMR also provides parent information, not only on medicine, but on behavioral strategy support groups and education interventions.

In addition, a second Nemours program, HRSA-funded program addresses this issue by placing clinical child psychologists in medically under-served areas in the State to provide assessment and psychotherapy services for these children.

Initial data from our integrated care program shows a 35 percent drop in diagnosis after one year. The rapid increase, the number of children, particularly younger children treated with medicine, does need to be examined closely. However, we're trying to limit what schools say to parents or frightened parents will not be sufficient control of the trend. Our State Task Force has determined the only solution is to develop comprehensive multi-modal treatment programs to address behavioral problems and encourage families to access and utilize all evidence-based therapies, not just medication as suggested by the Surgeon General's Conference in 2001. We, at Nemours, at attempting to do that for thousands of children we serve in Delaware and Florida.

Thank you.

DR. JOHANNESSEN: Thank you. Speaker number three, Dr. Lawrence Greenhill.

DR. GREENHILL: My name is Lawrence Greenhill. I'm a child psychiatrist working at the New York State Psychiatric Institute doing treatment research for the safety and efficacy of medications used to treat attention deficit hyperactivity disorder in children, adolescents and adults. I am the chairman of the Work Group on Research at the American Academy of Child and Adolescent Psychiatry. This organization strives to raise the standards of care and ethical methods in clinical trials and in medication. The organization also supports the FDA Advisory Panels and their efforts to better characterize the safety of medication treatments and we appreciate the broadcasting of tables, slides and all information on the website which helps practitioners and families.

This is important because untreated ADHD children can be highly impairing both for the patient and their family. Patients and families need to know about effective treatments for ADHD as well as full knowledge of their adverse events.

In order to help practitioners and their families, I'd like to pose three questions for the Advisory Panel to consider. First, how does one base clinical decisions on infrequent, but severe adverse events and do these infrequent events require new practices for the practitioner? I've seen the pendulum swing from the use of spontaneous patient interview methods which under-report the adverse events in small clinical trials to now the full disclosure of every adverse event from the unreliable post-marketing system, no matter how rare.

A full review of all infrequent, cardiovascular, neuropsychiatric events makes it difficult for families and physicians to determine the balance of benefit versus risk when seeking treatment. Professional organizations, including the American Academy of Child and Adolescent Psychiatry and the American Academy of Pediatrics, now publish practice guidelines for practitioners to set a standard for care. They ask practitioners to weigh all common adverse events when they're making their treatment decisions. How then do we weigh the infrequent and rare?

My second question is can the Panel recommend to the pharmaceutical industry and FDA to agree on standard methods for collecting safety information so that infrequent moderate and severe adverse events in the real world settings get more reliably detected.

My third question is that I would like to ask the Advisory Panel if they would recommend the FDA better specify which groups are at risk for specific adverse events and if that is known. In the last hearing which was the February 9th hearing, they voted by a narrow margin, 8 to 7, to recommend black box warnings. And that's quite controversial because they didn't specify whether it was for children or adults.

In closing, I'd like to add that my travel to this meeting has been supported by the American Academy of Child and Adolescent Psychiatry,

DR. JOHANNESSEN: Thank you. Next speaker is number four, Dr. Adelaide Robb.

DR. ROBB: I have slides. I'm here on behalf of the ACAP and the National Institutes of Neurologic Diseases and Stroke to talk about the safety data from an NINDS-funded study called Quantity and an ADHD Treatment that was conducted at Universities of Buffalo, Cincinnati, Pittsburgh and Rochester. Next slide.

I'm the head of the DSMB for that study. There were 122 children, ages 7 to 12, who were diagnosed with ADHD and no comorbid medical problems. Next slide.

They were either assigned to methylphenidate, clonidine, methylphenidate plus clonidine or a placebo. Next slide.

And they were titrated up to final doses over the first eight weeks in the study and left on maintenance dose for the second eight weeks in the study.

Next slide.

Cardiovascular parameters and side effects with the Pittsburgh Side Effects Rating Scale were collected. The mean dose of methylphenidate in the methylphenidate-only group was 30 milligrams. The mean dose of clonidine in the clonidine-only group was 0.24 milligrams of clonidine and then in combination they had .23 milligrams of clonidine and 25 of methylphenidate. Next slide.

And these are some of the more common side effects divided by whether patients had clonidine or no clonidine and the most common ones in the clonidine group were nervousness, sleepiness, fatigue and headache. Next slide.

And several patients in each of the groups had more serious SAEs meaning that the symptoms were present more days than not for most of the day over the course of a week. Two people withdrew early, one for chest pain and tachycardia on methylphenidate, although the physical exam and electrocardiogram were within normal limits and a second patient withdrew for a mild elevation in QTc of 440 milliseconds who was on combination methylphenidate and clonidine. The EKG showed left ventricular hypertrophy by voltage criteria only. Echocardiogram was normal. Next slide.