FOOD AND DRUG ADMINISTRATION

P R O C E E D I N G S (8:08 a.m.)

MR. JEHN: I am going to start the meeting with a brief COI statement, and then move on. Again, this meeting today was originally going to have a closed session, but today is totally public. That session has been canceled.

This brief announcement is in addition to the conflict of interest statement read at the beginning of the meeting on November 3, and will be part of the record for the Blood Products Advisory Committee meeting on November 4, 2005.

This announcement addresses conflict of interest for the discussions of topic II, on alpha-1 protease inhibitor products.

In accordance with 18 US Code Section 208(b)(3), waivers have been granted to Drs. Donna Di Michele and Katherine Manno.

Dr. William Duffell, Jr. is serving as industry representative, acting on behalf of all related industry, and is employed by Gambro BCT. Industry representatives are not special government employees and do not vote.

With regard to FDA's guest speakers, the agency has determined that the information provided by these speakers is essential.

The following information is being made public to allow the public to objectively evaluate any presentation and/or comments made by the speakers.

Dr. Mark Brantly is professor of medicine and molecular genetics and microbiology and Alpha-1 research professor at the University of Florida, Gainesville.

Dr. Hans Peter Schwarz is associate professor of medicine, vice president, global pre-clinical research and development, Baxter Bioscience, Vienna, Austria.

As guest speakers, they will not participate in the committee deliberations, nor will they vote. This conflict of interest statement will be available for review at the registration table.

We would like to remind members and consultants that if discussions involve any other products or firms not already on the agenda, for which an FDA participant has a personal or an imputed financial interest, the participants need to exclude themselves from such involvement, and their exclusion will be noted for the record.

FDA encourages all other participants to advise the committee of any financial relationships you have with the sponsors, products, competitors or firms, that could be affected by the discussion. Thank you.

DR. ALLEN: Thank you, Mr. Jehn. We will open the discussion this morning with an information update, serious adverse events following falsely elevated glucose measurements resulting from administration of an IVIG product containing maltose.

Presenters from the FDA will include Anne Gaines, Dr. Gaines, Dr. Pierce, and Ms. Bernhardt.

Agenda Item: Information - Serious Adverse Effects Following Falsely Elevated Glucose Measurements Resulting from Administration of an IGIV Product Containing Maltose.

DR. GAINES: Good morning. The topic of the presentation was just read to you. So, I won't bother reading that again.

I would mention, though, that this presentation has been divided into three segments. The first segment is an adverse event case report summary. The second and third segments will address other regulatory, biochemistry and device aspects of this issue.

Artifactual hyperglycemia -- and I will try to enunciate, perhaps sounding like an affectation, because we are talking about hyperglycemia and hypoglycemia, and obviously the distinction will be important.

Artifactual hyperglycemia, or falsely increased glucose results, may occur in patients who receive parenteral maltose, parenteral galactose, or oral d-xylose- containing biologic and drug products.

This artifactual hyperglycemia would most generally be manifested when blood glucose levels are measured with methodologies that are non-specific for glucose in the presence of maltose, galactose or d-xylose-containing products.

The clinical consequence of artifactual hyperglycemia is that falsely elevated glucose results, due to maltose, galactose or d-xylose, may result in life threatening or fatal hypoglycemia.

This life threatening or fatal hypoglycemia may result in two instances. The first instance would be when insulin is inappropriately administered for artifactual hyperglycemia.

Actual patient data from a patient who was receiving a maltose containing intravenous immune globulin product will serve to illustrate this point.

In a patient who was receiving this maltose containing intravenous immune globulin product, when the blood glucose level was measured using a glucose non-specific methodology, the result was reported as 231 milligrams per deciliter. This represented a falsely increased value, due to the presence of maltose.

When the patient's blood glucose level was measured using a methodology that was specific for glucose, the result reported was 84 milligrams per deciliter, and this value represented the patient's actual blood glucose.

So, had a treatment decision been made on the basis of the first glucose result, the 231 milligrams per deciliter, insulin might have been indicated.

However, based on the glucose result obtained using a glucose specific methodology, which gave the patient's actual glucose level, insulin would not have been indicated.

The second instance which may be associated with life threatening or fatal hypoglycemia is when treatment is not provided for actual hypoglycemia.

Again, data from a patient actually receiving a maltose containing intravenous immune globulin product can serve to illustrate this point.

When the patient's blood glucose level was measured using a glucose non-specific methodology, the result was reported as 167 milligrams per deciliter. Again, this represented a falsely increased glucose value due to the presence of maltose.

When the patient's blood glucose level was measured, however, with a glucose specific methodology, which reflected the patient's actual blood glucose level, the result was reported as 41 milligrams per deciliter.

Had the patient been assessed on the basis of the first blood glucose level, or the 167 milligram per deciliter level, no action would seemingly have been indicated.

However, if the patient was assessed on the basis of the second blood glucose level, which reflected the patient's actual blood glucose level, the patient would, as a minimum, needed to have been monitored, if not treated, for hypoglycemia.

CBER is aware of six case reports involving maltose containing intravenous immune globulin products that were associated with falsely elevated blood glucose measurements.

The first two case reports listed there were associated with fatal outcomes. The only case for which we have an appreciable amount of information is the first case. However, the other cases will also be summarized on the basis of what information was available.

The first case occurred in the United States and was reported to us in July of 2005. This case involved an 86-year-old male patient with a complicated medical history, including medical history of diabetes melitis.

He was admitted to the hospital with a four day history of cellulitis of his right foot, which rapidly progressed in a necrotizing fascitis and sepsis.

He received Octagam for treatment of septic shock. Octagam is an intravenous immune globulin product that was licensed by CBER in May of 2004.

It was initially licensed in 1993 in Europe, and is currently distributed worldwide. Octagam contains 10 percent maltose.

Septic shock represents an off label use of Octagam. The patient experienced a very complicated hospital course, including amputation of his right leg above the knee, renal dialysis, transfusion with various blood components and blood derivatives, among other medical interventions.

The patient's blood glucose levels were monitored with a glucose meter, a point of care device, that used a glucose non-specific methodology.

Similarly, his insulin doses were adjusted on the basis of the glucose meter levels which, again, were glucose non-specific methodology.

The patient became hypoglycemic, became comatose, was diagnosed with irreversible neurological damage, and expired following removal of life support measures.

This chart shows the patient's blood glucose levels in conjunction with octagam administration during hospital days three through five.

I will point out that the X axis represents the patient's sequential blood glucose measurements during this time frame, and it represents not an interval time line. It just is sequential data points. So, it is an ordinal time.

The blue horizontal line represents when octagam was administered. Octagam was administered as a five percent solution in three doses, as you can see, 10 grams, 70 grams, 35 grams respectively, that was infused at approximately 150 mls per hour.

The pink data points represent the patient's blood glucose levels that were obtained using a glucose specific methodology.

These pink data points represent the patient's actual blood glucose, regardless of whether or when octagam was administered.

The yellow data points represent the patient's blood glucose levels that were obtained using a glucose non-specific methodology.

Prior to the administration of octagam, these yellow data points would also reflect the patient's actual blood glucose level.

However, following the administration of octagam with increasing maltose concentrations in the patient's blood, these yellow data points would represent increasingly falsely increased glucose levels.

This graph is identical to the previous graph, with the exception that this shows a patient's blood glucose levels in conjunction with insulin administration.

The horizontal orange lines represent when the patient received regular insulin that was administered subcutaneously in doses ranging from two to 20 units.

The horizontal green line represents when the patient was administered IV drip insulin, and that was administered in doses ranging from 12 to 24 units per hour.

We can see that, as the octagam infusion began, the patient's blood glucose level started to rise. To appreciate the magnitude and the clinical consequence of these falsely increased values, I will just point out that the pink data points and the yellow data points representing the glucose specific and glucose non-specific values originally, initially, show a one to one correspondence.

For example, these represent values that may have either been simultaneously obtained, or obtained close enough in time that they can be reliably compared. You can see that the two values essentially reflect the same number of milligrams per deciliter. However, as the octagam infusion continued, you can see the pink and yellow data points diverge.

Here, and here in particular, you can see there is a marked divergence and, at the extreme, the difference between the patient's blood glucose values measured, using glucose specific and glucose non-specific methodology, showed a discrepancy of 270 milligrams per deciliter.

At the beginning of the octagam infusion, the patient was able to respond to verbal stimuli. As the falsely increased glucose levels were used to monitor his insulin doses, additional insulin was given and, by the time his blood glucose level began to decrease, his responsiveness to verbal stimuli also decreased accordingly.

At this point in time, indicated by the arrow, the patient was noted to be non-responsive. A blood glucose level obtained at this point in time, using the glucose non-specific methodology, revealed a blood glucose level of 115 milligrams per deciliter.

The corresponding pink data point, representing the blood glucose level using the specific methodology, was 12 milligrams per deciliter.

The patient's responsiveness never changed throughout the remainder of his hospitalization, up until the time of his death.

The second case report involves a case that was reported to the medicines and health care products regulatory agency in the United Kingdom in 2002.

This case involved a 50-year-old diabetic male with a failing pancreas renal transplant, who received octagam as an anti-rejection agent. This represents an off label use of octagam.

Based on artificial hyperglycemia, as a result of the maltose contained in the octagam, the patient had inappropriate insulin administered.

He became severely hypoglycemic, became comatose, was diagnosed with severe nervous system depression, and expired. No further information was provided.

The third and fourth cases represent cases that were reported to the French health product safety agency in 2003. Both of these cases involved three-week-old male, non-diabetic patients, who received octagam for unknown indications.

Both of these patients developed artifactual hyperglycemia, although this was apparently recognized and resolved without the need for medical intervention. However, we don't have any additional details on these cases.

The fifth and sixth cases were reported in the Medical Journal of Australia in 2004. Both of these cases involved an intravenous immune globulin product by the name of Intergam, which is not licensed in the United States. Intergam-P, though, does contain 10 percent maltose.

The first patient was a 64-year-old diabetic female on dialysis, who received intergam-P for ITP, or immune thrombocytopenic purpura. This represents a labeled indication for this product.

The patient subsequently developed artifactual hyperglycemia, for which insulin was administered in increased doses.

The patient then developed hypoglycemia, which resolved without apparent sequelae. Again, no further details were really provided.

The second case was a 35-year-old non-diabetic female on parenteral nutrition who, likewise, received intergam-P for ITP.

She, likewise, developed artifactual hyperglycemia, which was recognized and apparently resolved without the need for medical intervention. Again, there is limited information that was provided on this case.

We, unfortunately, do not know how frequently artifactual hyperglycemia occurs. We, more important, have no idea how often it occurs associated with serious adverse events.

The reason why we don't have answers to any indications of frequency are that FDA's adverse events surveillance for licensed products uses primarily passive surveillance.

There are numerous limitations to passive surveillance, and FDA's medwatch or adverse event reporting system is an example of a passive surveillance system which serves as our primary source of data for adverse events.

Among the various limitations of passive surveillance systems, including medwatch, is the reliance on voluntary reporting of adverse events by physicians, other health care professionals, patients, and others.

It also requires that someone suspecting that there is a possible causal relationship between a product and an adverse event, there are various other factors that influence this.

Newly licensed products tend to have more reports submitted than older products. Publicity, whether it is in the newspaper or some sort of presentation, may prompt what we call stimulated reporting.

Whatever the limitations are, though, the net result is that we are confident that we have a significant amount of under-reporting.

However, the extent of the under-reporting is unknown. It probably varies from product to product. .So, there is no way we can correct our data, so to speak, to reflect some sort of frequency estimate.

So, even incidence rates or estimated reporting rates, under normal circumstances, without additional data sources, cannot be calculated from our data.

The limitations of passive surveillance, specifically the Medwatch adverse event reporting system notwithstanding, Medwatch has enabled us to detect, investigate and act upon numerous adverse events.

For those reasons, we encourage physicians, other health care professionals, patients and others, to report adverse events, particularly serious adverse events, to FDA for any FDA improved product, including the maltose-containing intravenous immune globulin products.

Adverse event reports can be submitted directly to FDA by internet, by telephone, by fax or by mail. Instructions and forms for submitting adverse event reports to FDA are listed at the cited Medwatch web site.

Alternatively, adverse event reports can be submitted to FDA through manufacturers or distributors, and I say to FDA, because manufacturers and distributors, in turn, are required to submit reports to us that they receive about adverse events.

Contact information for reporting adverse events to manufacturer or distributors is generally available in professional package inserts, or on their sponsored web sites. I thank you for your attention.

DR. ALLEN; Thank you very much. We will move on to the other presentations, and then have time for questions for all the speakers at the end. Dr. Ross Pierce?

Agenda Item: Information.

DR. PIERCE: Good morning. I am going to talk a little bit about the different point of care glucose testing systems, and also the CBER response to this problem that you just heard about.

So, since their introduction, point of care glucose meter devices have become widely used, not only for home blood glucose monitoring, but also in hospitals and clinics.

The test strips that use the glucose d-hydrogenase pyroloquinoline quinone, or GDH-PQQ method, are not specific in every case for glucose, in that maltose can be falsely read as glucose by that enzyme.

At this time, we believe that all of the test strips that use the GDH-PQQ test method are, in fact, labeled with some type of precaution regarding the potential for maltose in the patient's blood to cause falsely high readings of blood glucose.

Non-specific glucose test methods, as you have just heard, are unfortunately, from time to time, still inappropriately used in patients who have received maltose-containing products, as reflected in the horrific case that we heard about from Dr. Gaines.

There are actually two IGIV products licensed in the United States that contain maltose and have this potential.

The difference in the various test kit systems include different enzymes as well as different glucose d-hydrogenase enzymes.

There are two of those that are used in test strips for glucose marketed in the United States, and the GDH-PQQ test method does not use nicotinic acid adenine dinucleotide, NAD, as the enzyme cofactor. It uses PQQ as the co-factor.

This enzyme uses either glucose or maltose as a substrate. Many test systems employ this GDH-PQQ method that is subject to maltose interference.

In contrast, another glucose d-hydrogenase, the NAD dependent glucose d-hydrogenase, uses glucose but not maltose as a substrate. So, maltose will not interfere. This is used in at least one glucose meter test strip system licensed in the United States.

This slide depicts the biochemical, chemical and electrochemical reactions that occur during the assays that are GDH-PQQ based.

As you can see, there is a redox reduction and maltose or glucose is converted into the corresponding lactone by the action of the GDH-PQQ enzyme, and the change in the redox state of the co-factor translates into an electrical signal.

There are other GDH-PQQ test methods, where the readout signal is a calorimetric method, but the enzyme is still subject to maltose interference, because it uses the GDH-PQQ form of glucose d-hydrogenase.

This shows that the co-factor PQQ is structurally distinct from the co-factor used by the other enzyme that is not subject to maltose interference, NAD.

Here you see that the reactive center in glucose is the same structurally as the reactive center in maltose, in terms of the action by the GDH-PQQ enzyme.

So, if we divide these various test strip systems into those that are subject to maltose interference on the left, and those that are not subject to interference by maltose on the right, on the left, we only have DGH-PQQ at the present time.

Now, most of the point of care test strips that are used with glucose meters, that use this enzyme, name the enzyme as PDH-PQQ, but there is at least one point of care calorimetric system that terms this same enzyme a glucose di-oxidoreductase, which is really a translation of the name of this enzyme from the German.

Now, one of the advantages of the GDH-PQQ -- you have already heard about the disadvantage, that it falsely interprets maltose as glucose -- but it does have an advantage, in that it is not subject to a skewing of the readout by low oxygen tension in the patient's blood.

If we go over onto the right hand side, you can see that the NAD dependent GDH enzyme does not interfere, as I mentioned. This is also not subject to interference in the readout by low oxygen tension.

Other more specific methods are hexokinase, where low oxygen tension can interfere with the glucose reading, and glucose oxidase, which is also subject to a bias by low oxygen tension.

Also, mannitol in extremely high concentrations, can interfere with the glucose oxidase methods, which is one of the popular clinical laboratory based methods for glucose determination.

So, which are the IGG products that contain maltose? Well, in terms of polyclonal IGG products, non-hyperimmune products, we have two in the United States, octagam five percent licensed by Octapharma, and gamimmune-N five percent, which is no longer actively distributed in the United States, once its sponsorship was transferred from Bayer to Talecris. However, we believe there may still be some product bearing the Bayer label still in the market place.

Another maltose containing immunoglobulin that is a hyperimmune globulin is WinRho SDF liquid, marketed by Cangene. However, at the label doses, we do not expect that the maltose concentration will be high enough to interfere with glucose because of the comparatively low dose of that product. Off label use might be a different story at higher doses.

Vaccinia immune globulin was approved this year. The license holder is Cangene. This is the product for complications of vaccinia vaccination. It is only available through the CDC, and interference is expected at label doses of that product.

I should also mention that, while WinRho SDF liquid is an approved product, its marketing launch hasn't yet actually occurred.

So, what are the actions that CBER has taken in response to these adverse events? Well, as soon as we found out about this most recent case, we quickly formed a CBER CDRH working group.

We worked to strengthen the labeling of all the maltose containing IGIV and immunoglobulin products. We asked the sponsors of the IGIV products to issue important drug warning letters to physicians, customers and hospitals, and one such letter has issued as of this time.

There has been coordination between -- extensive coordination between -- CDER and CDRH on health alerts for FDA web sites and Medwatch list serves regarding this problem, and we are in the process of drafting an article to be submitted to one or more widely circulated medical journals.

The work that we have done with the sponsors regarding the package inserts is to add a warning -- none of the products had this in the warning section previously -- and to strengthen the precaution section.

In the sort of class labeling that we have developed for the maltose containing IGIVs, the new labeling will indicate that some types of blood glucose test systems, such as GHPQQ or glucose dioxidoreductase methods falsely interpret maltose as glutose, that this has resulted in inappropriate administration of insulin, resulting in life threatening hypoglycemia and that, when administering maltose containing IGIVs, use a glucose specific method for measuring glucose, and instructs the users and clinicians to carefully review the product information of the glucose test system, including that of the test strips, to determine if the system is appropriate for patients receiving maltose containing parenteral products.

Some of the products only contain the information about maltose interference in the test strips, and not in the package insert of the glucose meter itself. Contact the manufacturer of the test strip if there is any uncertainty.

So, we would like to acknowledge the tremendous amount of cooperation that we received from the hospital pharmacist and the physician who cared for the patient that you heard about from Dr. Gaines this morning.

The case that she described was recently published in an ISMP safety alert on September 8 with an erratum in the next issue. Thank you.

DR. ALLEN: Thank you very much. The third speaker from the FDA will be Ms. Patricia Bernhardt.

Agenda Item: Information.

MS. BERNHARDT: Good morning. I am going to present the CDRH perspective on this issue. CDRH regulates the test strips that are involved in this.

The manufacturers of these test strips, as part of their evaluation of the devices, prior to marketing, evaluate many types of possible interferences, and sugars are one of the interferences that they do look at.

The types of sugars that have been evaluated by the manufacturers are maltose, galactose, xylose, lactose, and sorbitol.

Maltose has been evaluated. I don't have the concentrations -- the highest concentrations -- that the manufacturers have evaluated.

What they have found is that, at concentrations greater than 13 milligrams per deciliter, the interference begins to be seen.

For galactose, the interference is seen at greater than 10 milligrams per deciliter. For d-xylose, the interference is at greater than nine milligrams per deciliter, and d-xylose is typically administered during a xylose malabsorption test and, during that test, the blood xylose concentration reaches approximately 30 milligrams per deciliter.

Because of that, the xylose interference, the warnings in the labeling are that these glucose test strips should not be used during the course of a xylose malabsorption test.

Sorbitol, there is no interference up to 70 milligrams in the blood, and lactose, there is no interference up to five milligrams, and that is 10 times the normal concentration in the blood, which is 0.5 milligrams per deciliter.

CDRH has been working extensively with CBER, as Dr. Pierce mentioned. We are working to post a reminder on the OIVD web site, and the FDA diabetes web page about this interference with these sugars.

These interferences are listed in the warnings in the test strip inserts for the glucose d-hydrogenase PQQ enzyme. We are working with the manufacturers to strengthen those warnings, and to make them more consistent. There is not much consistency between the different labeling for these products.

We are also developing an upcoming patient safety news item, which will talk about this interference. As I have mentioned, we have requested labeling modifications to include the maltose, galactose and the xylose interference, and we are requesting it to be in a more consistent fashion.

An example of a warning from the GDH PQQ test strip label is up on the screen. That is basically the format that we are going to be hoping to see this warning in the labeling in the future.

In addition to the labeling warnings, which have been in the labeling for quite some time, there is a Clinical Laboratory Standards Institute document on glucose monitoring settings without laboratory support, that does mention this maltose interference.

It does not mention galactose or xylose, but it does mention maltose. It lists maltose as a cause of falsely elevated results with some glucose d-hydrogenase systems.

So, the information is out there. We hope that we can increase the health care provider's awareness of this interference so that, when they use these glucose testing methods on their patients, and are administering some of these products that contain some of these sugars, there will be heightened awareness of the potential interference that can be caused. Thank you.

DR. ALLEN: Does that end the FDA presentations? You are well under your allotted time for that. Questions from the committee rom any of the speakers? This is just an informational item. We are not being asked for any advice.

Agenda Item: Questions from the Committee.

DR. DUFFELL: When the medications were being used, were they used on label or off label?

DR. PIERCE: In the two cases that resulted in fatalities, or were associated with fatalities, those were off label uses of the IVIG products.

DR. DUFFELL: In the other cases, you don't know, then?

DR. GAINES: The first two cases, as Dr. Pierce mentioned, were off label use. The third and fourth cases, the French cases, the indication for use was not specified, was not provided, so we don't know.

The fifth and sixth cases, which were from the Australian journal article, were for on-label use for treatment of ITP. Those latter two cases involved an IVIG that was not licensed in the United States, although it was a 10 percent maltose containing intravenous immune globulin product.

DR. CRYER: It is now very commonplace in intensive care units, particularly septic patients or severely injured patients, to be monitoring glucose in a very narrow range using an algorithm between 80 and 110 millimolar.

Obviously a nurse, without a physician, is using an algorithm to give insulin to keep it in that very narrow range.

It seems to me that the important thing to find out would be how are they measuring the glucose in those protocols. I honestly don't know.

I do it in my own ICU, but I am not sure whether they use the values by the lab or a glucometer at the bedside. Do you have any information, across the country, what ICUs are doing to use those protocols?

DR. PIERCE: I don't, but there is obviously a great temptation to want to use the point of care device, because of its virtually instantaneous feedback. It obviously takes a longer period of time to get a stat glucose back from the clinical lab.

So, there is pressure to move in that direction, but people have to be cognizant of the potential down sides from interfering medicines.

Also, in sepsis, there is a difference between the capillary blood glucose level and the venous blood glucose level. So, the venous blood glucose level can be significantly lower than that of the capillary blood glucose level that is measured typically with the point of care devices. This needs to be taken into account in these protocols as well.

DR. KULKARNI: I was wondering, if there a dose dependent effect? We use IVIGs for all different doses for ITP indications.

DR. PIERCE: We believe that the lowest recommended dose in the labeling of any IVIG product that contains maltose is capable of producing this false elevation of glucose measurements using the GDH PQQ method.

DR. SIEGAL: The dose that was given in the U.S. case was something on the order of 60 grams over the time that he was given gammaglobulin?

DR. GAINES: he received three different doses, or three different administrations. The first one was a 10 gram dose, the second was a 70 gram dose. The third one, which actually was after the fact, so to speak, was a 35 gram dose. This was a 63 kilogram patient.

DR. SIEGAL: So, about 1,000 milligrams per kilo would be the dose roughly that he got. I mean, you wonder about hyperozmolarity in this setting. Obviously, it is pretty short lived, and people use huge doses of gammaglobulin and ITP in other settings that may really create problems for patients that are real, not just fictitious in the sense that there is a lot of disaccharide floating around for the time being. Has that ever been considered an issue?

DR. PIERCE: The osmolarity of IVIG products has been considered when evaluating the potential for adverse events, and it is something that we always consider when looking at new products that are coming to use for investigation.

DR. KLEIN: I just want to make sure that I understand correctly. Is there no point of care test that is specific for glucose, or is it that the specific ones with glucose oxidase have other disadvantages, such as sensitivity to low PO2?

DR. PIERCE: Maybe Dr. Bernhardt can respond to that.

MS. BERNHARDT: Yes, there are several that use the NAD version, and there are many that use the glucose oxidase version.

It is the hospital's choice which one they choose to buy. My personal opinion is that the people responsible for purchasing the equipment to be used in these point of care settings should be more aware of these limitations for these devices in their particular settings, based on the types of medications and treatments that are used for the patients.

DR. KLEIN: At the risk of sounding simplistic, is the licensure of something that is non-specific at point of care, is this historical? Is it much cheaper to buy and, therefore, a better thing to do, or is there some reason maybe to think about not having something that is non-specific that is on the market?

These are just a couple of the drugs that contain interfering substances, and clearly there will be more in the future.

MS. BERNHARDT: The main purpose for these devices is not for point of care use. It is for home testing by diabetics to monitor their individual glucoses, and the hospitals choose to sometimes use these in point of care settings, as Dr. Pierce said. It is probably quicker than waiting for a stat result from a laboratory, but that is not the real intent of the devices.

DR. KLEIN: Are they cheaper? Is there only one manufacturer of specific ones?

MS. BERNHARDT: No, I don't believe so.

DR. ALLEN: That is a very interesting issue in terms of devices developed and intended for one use not being restricted in any other way.

DR. NELSON: I can give a personal account that it is not only the ICU, but post-surgery patients also have glucose monitored insulin given, and it is a rapid device. It happened to me six months ago.

DR. KUEHNERT: I had some questions about surveillance. You mentioned the passive surveillance aspect. Was the U.S. case detected through Medwatch? I am just curious how it came to attention by the hospital.

DR. GAINES: It is a long complicated story, and I am not good at telling short stories, but I will do my best. We became aware of this particular case initially because of a list serv.

I am not enough of a computer person to really understand what list servs are, but it was a specialized medical specialty e mail list serv, where this case was reported.

The queries were, has anyone seen a case like this. The physician and pharmacist had actually worked up the case reasonably thoroughly to know what the problem was.

The report has since been reported to Medwatch, after I investigated the case along with the hospital personnel as a pro forma matter. It is in Medwatch, but that is not how we became aware of this one.

DR. KUEHNERT: There are a lot of ways that these things come to attention, and Medwatch is only one. So, I would just encourage you to think creatively about how to improve surveillance on this.

I don't know how many hospitals look when there is an adverse event associated with someone who is on insulin, if they look to see what other products they received, but that would be a good idea to maybe track that.

Just along the comments by Dr. Klein, I don't know how many other products contain these sugars to know if there are other events that are occurring to those products as well. That would be something certainly of importance.

In trying to get the message out, you mentioned the list servs and other ways to do that, but I was wondering, when you were talking about labeling of the strips, I saw this letter. Is there also a plan to put a warning on these products, and also, looking at these other products that have these sugars as well?

DR. GAINES: Are you referring to the test strips or the intravenous solutions?

DR. KUEHNERT: The intravenous solutions. I think it mentioned the test strips, but I wondered what about the solutions?

DR. PIERCE: The IGIV products, the labeling of all of them is in the process of being revised, to reflect something in the warning section as well as a subsection of precautions on drug laboratory test interactions.

We are continuing to work with CDER and other parts of the agency to try to identify other products which might interfere with the GDH PQQ testing method, and we are hopeful that the other centers that have purview over those products will take appropriate action.

DR. KUEHNERT: Just a last question. When something like this happens, is there some sort of pharmacy action that occurs, that when you have insulin and a product like this, that there is some sort of warning generated? Do you know if that is something that happens in hospitals?

DR. PIERCE: It is an excellent question. I think it should happen in this type of case, and pharmacies do, as you point out, now more and more, particularly with the electronic sort of ordering systems, it is very easy for these automatic potential drug drug interaction warnings to come up. So, this is certainly something that would be good to consider instituting.

DR. ALLEN: I think to answer your question from a different perspective, if you have been or worked in one hospital, you have worked in one hospital.

There is no any degree of uniformity. It would depend on the awareness of the hospital pharmacy system and the working committees, the medical care structure within the hospital.

We have got such a complex -- I don't want to use the word system. We have got such a complex structure in our country that there isn't any single way to do that.

The issue of surveillance or reporting, you have to appreciate or recognize the adverse event for what it is before it can be adequately reported.

Then people have to know that the reporting system is in place. As you pointed out, the degree of under-reporting with the existing mechanisms is probably fairly significant.

I think many hospitals today, certainly when I was in training, which was a number of years ago, we had regular mortality review conferences. A high percentage of patients in hospital were sent for autopsy, we had lots of adverse review settings.

I think that is much less common today, even in university settings, and certainly at the community level hospital setting. I don't think there is nearly as much post-event review as would be very helpful.

DR. KUEHNERT: Just to make a follow up comment to that, other than in Medwatch, I am not aware of product surveillance for adverse events for IVIG or other products. So, that is something I guess we can think about in the larger picture along with transfusion adverse event surveillance, but it is just lacking in the United States.

I don't know, in other countries that have hema-vigilance programs, whether they include other products, but it is certainly something that will be of interest in the future.

DR. KLEIN: The U.S. Pharmacopoeia has a system, but I don't know whether it would cover all hospitals.

DR. CRYER: Just one other comments about a solution. In fact, the data that was provided indicated that the hospital at least was collecting the right information.

So, I guess if you are going to use point of care testing, the advantage of that is efficiency. You know right away rather than having to wait for the lab test to come back.

Then you probably ought to have duplicate laboratory tests with some sort of warning system as soon as the lab knows that you have a value outside the range, which almost all hospitals have.

You would think that there would be some warning in some reasonable lag time that this was occurring. It seems to me that a solution is just to make sure that that duplicate testing is being done when doing point of care testing.

DR. PIERCE: That is certainly one approach. In this case, it was obvious form a review of the data that the hospital personnel were relying primarily on the point of care glucose meter.

The frequency with which laboratory-based glucose measurements were being obtained dropped off dramatically in the last day before the patient really deteriorated in terms of his neurological status.

If that blood sugar of 12, if that had been obtained six or 12 hours later than that, that could also have affected the timing, the recognition of the problem.

DR. ALLEn; Dr. Cryer's point, however, might be something that could be included in the discussion section of the article. I think that is a very important warning.

DR. STEFANO: If I could address one of the concerns that was just raised, next week in Chicago, we are actually just starting to find out, through the ICH -- the International Conference for Harmonization -- how many different systems that are in place in terms of notifications.

So, next week, on the table for discussion, is to hear what the EU and the Japanese and so on, all the members, are doing in terms of notification, so that we understand what everyone is doing around the world, and try to find a point at which we can harmonize, if that is even remotely possible. So, that might answer some of your concerns.

DR. KUEHNERT: Great, thanks.

DR. DI MICHELE: I just wanted to add that, certainly among blood product derivatives and clotting factors, I think there has been a move lately to encourage post-licensure pharmacosurveillance in a much more organized way, whether it is industry based or non-industry based.

I would like to add my vote for that as well, because almost everything we learn about products, we learn post-licensure.

DR. QUINN: Jim, you may know more about this, but would this be construed as medical error? The American Association of Accreditation of Hospitals is trying to standardize and get a reporting system in.

Now, a lot of hospitals would probably not want, for the points that you issued, they would not want a lot of this, because it becomes public information.

Nevertheless, that is somewhere where FDA could interface with the American Association for the Accreditation of Hospitals, for this type of reporting system.

I think physicians, health care providers, need to be made more aware of this interaction. For us working on wards, we give massive amounts of intravenous gammaglobulin for ITP for many of our patients, and this kind of information is not as readily available.

Pharmacies need to alert it if you are using this. When a patient is diabetic, you need to be aware of what to use to monitor it and so forth.

DR. ALLEN: You ask, do I know, and I don't, but I think it is a very good point and one that is worthy of additional follow through.

DR. SIEGAL: Both as a user of a glucometer and also as a purveyor of IV gammaglobulin a fair amount of time, I was sort of vaguely aware of the hyperglycemia issue because I have seen it actually, and I assumed that it was not real.

In fact, I think people in medicine and nursing probably are blissfully unaware of the shortcomings of glucose monitoring.

I am sure that most people who actually use these things have no idea that PO2 and other medications could interfere or, if they are aware, it is only subliminally.

It ought to be stated, and maybe directed at the people who are likely to do this -- intensivists and hematologists and immunologists and people who use these products a lot.

DR. ALLEN: I certainly think that the steps that have been taken are the right ones, but clearly a lot of information, disseminated broadly to different groups within the medical and hospital care setting needs to be done as part of this.

DR. KULKARNI: I just wanted to know, how long does the hyperglycemia last? The immunoglobulin lasts for a long time in the body.

I am just wondering, once these patients go home, do they continue to be aware of this glucose monitoring and the problems with it?

DR. PIERCE: The duration of this is going to be dependent on a lot of different factors and is going to vary form patient to patient.

The position that we have taken is that, if these products are used -- the maltose containing products, for example, in a particular individual, that they just should be using an alternative method for the determination of glucose.

We are not in a position to say, well, you can go back to using the GDH PQQ method after X number of hours or days.

MS. BAKER: Has the American Diabetic Association or any other patient advocacy organizations been engaged in discussions about this?

DR.PIERCE: The important drug warning letter issued by Octapharma was sent to the diabetologists from the mailing list of the Endocrine Society. We have not had direct discussions with the organization that you mentioned, however.

DR. DUFFELL: Judith's question actually picks up on the thought that I had. We were talking yesterday about labeling and putting a lot of things in labeling.

This is another case where, since really the intended audience for most of these devices is home use -- it is not really point of care -- I am wondering if we have done a good study of that warning and the impact it is going to have on the end user, whether it is going to cause undue concern, and whether or not it even means anything to them, as it is worded here.

I don't know if most patients would understand what the significance of this is. So, coming back to Judith's remark, it may be a good idea to try to test pilot, with the appropriate organizations, how most end users actually perceive this type of warning. Is it achieving the effect, or is it even appropriate for them.

If it is point of care where this is occurring, maybe you are going at it the wrong way to put something into the device labeling, and it should be some other means of communication.

MS. BERNHARDT: These warnings have been in the device labeling for quite some time. For over-the-counter products, which these glucose meters and strips are primarily considered to be, we do consumer testing of the labeling to ensure that it can be understood at a seventh grade level.

The labeling regs that we are under require certain other information to be in the labeling, information that a typical lay user may not understand, but it is required by the regs to be in there.

Some manufacturers choose to have a completely separate labeling for professional users versus lay users. Most of them, however, just have the one labeling, and this information is usually in a separate section that is addressed to health care professionals.

I think the diabetes nurse educators, that do work with a lot of the diabetics, they are very aware of the limitations for these devices, but the hospital personnel may not be as aware.

DR. DUFFELL: So, the warning that was put up in that slide already exists. This is not a fix. This is something that is already being done.

MS. BERNHARDT: That is actually from a label. That was an example from a label that we consider to be a complete warning for this particular interference, and that is what we are going to aim to try to get all the manufacturers to conform to that.

DR. DUFFELL: So, it is in one, but you are going to get all of them to use it.

MS. BERNHARDT: Yes, some of the other devices have warnings for just maltose, some of them will have a warning for maltose and galactose, but that particular one that I provided the example for is a complete one that addresses all those sugars, and that is what we would like to see in the future.

DR. ALLEN: Although, as was noted in the discussion yesterday, labels sometimes aren't read or followed.

I think the other part of that, that you raised, in terms of the point of care use for an instrument that wasn't designed for that, one wonders whether or not CLIA should be involved in looking at how point of care lab tests are done in hospital settings, which I assume would include certainly emergency departments, intensive care centers, perhaps surgical suites themselves a wide variety of settings in the hospital, where they may be used in ways that they weren't designed for or intended.

MS. BERNHARDT: Glucose meters on test strips are automatically waived by CLIA.

DR. ALLEN: I understand that, but maybe within a hospital setting, that waiver is not appropriate. That was the point I was trying to make.

All right, further discussion or questions from the FDA on this issue? That was a fascinating discussion. I am going to take the chairman's prerogative now.

Octapharma has asked to be able to make a statement on this topic. Normally this would occur during an open public hearing. We do not have an open public hearing scheduled at this time It is later in the morning toward the end of our discussion of topic two.

Given that the Octapharma statement is pertinent to this discussion, I will allow the Octapharma people to make a five-minute presentation at this time if you wish, or you may be invited to speak during the open public hearing. It is your discretion. Would you introduce yourself, please, and then you have got five minutes for your statement?

Agenda Item: Statement by Octapharma.

MR. SCHLEISS: First, I would like to thank the board for allowing us to give this presentation. Obviously, we have been actively involved in this process.

My name is Tom Schleiss(?). I am the medical affairs manager for Octapharma. Also, with me in the front row, is my esteemed colleague, Dr. Stephan Haag, who is the international business manager for Octapharma.

This interaction is a known characteristic of certain point-of-use glucose monitoring systems. As you have already heard, it is the GDH PQQ system that is being used that is non-specific for glucose.

This involves about 25 percent of the meters that are currently on the market. That is one statistic. The other statistic is, in the hospital setting, approximately 60 percent of the hospitals that use these point of use meters use a meter that does have this interaction. So, it is certainly a high percentage in the hospital, but in overall point of use systems, it is only about 25 percent of them.

Maltose, maltatriose and, to a lesser extent, maltotetrose, are misinterpreted by these test systems as glucose, and this results in this elevated glucose reading.

As we have seen in this one particular case, the falsely elevated glucose readings can result in the inappropriate administration of insulin, which can cause life threatening hypoglycemia.

The flip side of the coin is, a patient may be hypoglycemic but, because of the elevated reading, they may look to be in a normal state and, as a result, the hypoglycemic state may go untreated because of the falsely elevated glucose readings.

Maltose itself is not a dangerous drug. It has been used in a variety of products for years. It has been used in IVIG solutions, as is the case in ours. A similar type of maltose is used in peritoneal dialysis fluid, and that is where this interaction first became known, was in peritoneal dialysis solutions, and the interaction with the glucose monitoring systems which had the GDH PQQ systems. This was approximately 2003 that this came to light.

As a result, our product, and other products, have this labeling regarding this interaction in their package inserts. There is also at least one oncology drug that has maltose as part of its preparation.

The warning statements regarding this non-glucose specificity are normally present in the package inserts, specifically the test strips that are used with these glucometers. Information is present in there.

Warning statements regarding the potential for falsely elevated glucose readings with certain glucometers are normally present in the prescribing information of maltose, and icodextrin containing drugs, icodextrin being the peritoneal dialysis solutions. Icodextrin is converted to maltose, and then this interaction takes place.

Octapharma has had information regarding this interaction in our package insert since the introduction of our products into this country.

We have been in the U.S. market for about two years. We have been in other markets throughout the world for a significantly longer time.

We worked in collaboration, and under the guidance of the FDA and, as you saw there, we made changes not only to our warning but also to our precaution statements and our package insert.

We have also sent a dear doctor warning letter to the parties that were agreed upon between Octapharma and the FDA. I think you have already seen the language that we have added to our package inserts.

I also worked very closely with the Institute for Safe Medical Practices. They are another sort of a Medwatch type of an agency.

I personally worked very closely with them regarding this interaction and what happened with this patient and, as a result, they were really the first to post a very balanced piece of -- a summary piece which identified the interaction, looked at ways to avoid the interaction. It was a very, very helpful publication that went out to all pharmacists, and a fair amount of nurses also subscribed to this.

Dr. Haag and I have also worked very closely with Roche, who makes the Accucheck blood glucose monitoring system, which is the system that was involved in this particular case.

They are as eager as we are to help strengthen the language in their package insert, and help identify their hospital systems and their patients who are using these systems, of this potential interaction. Are there any questions?

DR. SNIDER: Have you spoken to nurses organizations about the interaction? In hospital use, nurses are really the people who most frequently use these tests, with respect to the physicians or the pharmacists?

MR. SCHLEISS: The information has gone out to practically any practitioner who would be involved on this. We focused specifically, in terms of pharmacists and nurses, on our clients, which would include everyone involved in their agencies.

We only use two distributors. Both of them received the information to pass along to their customers. So, our intent was, knowing that pharmacists and nurses are involved, our intent was to notify everyone.

Even from day one, this was an interaction which our sales force was instructed to inform all of our customers that this was present. This was nothing that we ever tried to hide.

In this particular case, it was a pharmacist who had not actually seen any of our representatives. So, he wasn't aware of it.

DR. ALLEN: Other questions or comments? Okay, thank you very much. We will move now into presentations and then discussion of topic two, the heterogeneity of commercial alpha-1 proteinase inhibitor products, human, the implications for longer-term safety and efficacy.

The first presentation, by Dr. Andrew Shrake of the FDA, will be the introduction and questions to the committee.

Before we proceed, I just wanted to acknowledge that Dr. Snider is sitting at the table as a temporary voting member for the topic two discussion. Welcome.

Agenda Item: Heterogeneity of Commercial Alpha-1-Proteinase Inhibitor (Human) Products - Implications for Longer-Term Safety and Efficacy. Introduction and Questions to the Committee.

DR. SHRAKE: Good morning. Topic two today is heterogeneity of commercial alpha-1 proteinase inhibitor products, implications for longer-term safety and efficacy.

Before starting our discussion of the heterogeneity of licensed alpha-1 PI products, I would like to briefly summarize the role that alpha-1 PI plays physically, ramifications of alpha-1 PI deficiency, and treatment of alpha-1 PI deficiency.

Individuals who are homozygous for mutant alpha-1 PI -- and the most common mutant is the Z mutant -- these individuals have low circulating levels of alpha-1 PI due to the tendency of Z mutant to fold more slowly during translation, giving rise to the formation of polymer which, in turn, gives rise to the formation of inclusion bodies within hepatocytes, which are, in fact, the major site of synthesis of alpha-1 PI.

Alpha-1 PI deficient individuals have chronic obstructive pulmonary disease, and usually present with progressive, ultimately fatal, emphysema, due to the action of uninhibited elastase in the lower lung, particularly on elastin, and some of these individuals manifest chronic liver disease, including hepatocellular carcinoma.

Such deficient patients receive alpha-1 PI augmentation therapy to slow the rate of progression of the emphysema.

At present, there are three plasma derived alpha-1 PI products for augmentation therapy of patients with congenital alpha-1 PI deficiency.

Prolastin was licensed by Bayer Corporation in 1987, and I believe just this past year Talecris bought the license.

Aralast was originally licensed by Alpha Therapeutic Corporation in 2002, and shortly thereafter Baxter purchased the license. Zemaira was licensed by Behring in 2003.

Augmentation therapy involves weekly intravenous administration of a licensed product at a dose of 60 milligrams of functionally active alpha-1 PI per kilogram of body mass, in order to maintain circulating levels of alpha-1 PI at 11 micromolar or greater, as measured antigenically.

For a 75 kilogram person, this corresponds to infusion of 4.5 grams of alpha-1 PI per week, throughout the lifetime of the patient. That is a lot of protein.

I would like to give a brief overview of the events that gave rise to the issues that are under consideration this morning.

On September 16, 2004, Dr. Mark Brantly from the University of Florida Medical School, contacted our laboratory concerning an atypical isoelectric focusing pattern that he had observed in analyzing sera of patients who were on augmentation therapy with aralast.

Then, on September 24, CBER initially contacted Baxter to request lots of aralast for isoelectric focusing analysis, both by CBER and by Baxter.

Then, on October 14, 2004, the Alpha-1 Foundation contacted the CBER center director's office reporting the same observation -- that is, the atypical IEF pattern for aralast, but the report was from a different laboratory.

On October 15, CBER contacted Baxter again and, at that time, requested pharmacovigilance data, and inventory samples of both the product and patient serum samples from clinical studies, Baxter isoelectric focusing results, and at the same time also requested that Baxter initiate an investigation into the root cause of the atypical IEF pattern for aralast.

During a telecon on October 20, 2004, Baxter speculated to the FDA that the additional isoelectric focusing bound is, in fact, a rare glycoform that becomes highly enriched during the purification of aralast.

In contrast, during that same telecon, CBER contended that there was an anodic shift by a whole negative charge of most of the major BADs. At least at that point we felt that was a very reasonable explanation.

We also communicated to Baxter during that telecon was that our IEF results also demonstrated consistency of manufacturing back to the conformance lots that were used for the licensure of aralast.

During November 2004, CBER continued the investigation by comparing isoelectric focusing properties of all three products with those of alpha-1 PI in plasma, and then reported the results to CBER management on December 7, 2004.

Shortly after that report, CBER confirmed the consistency of the atypical IEF properties of aralast during the phase III clinical trial which was used for licensure of aralast. These involved data that I believe Dr. Brantly is going to talk about later.

On January 11 of this year, the Alpha-1 Foundation sent patients undergoing therapy for alpha-1 PI deficiency a letter notifying them of the atypical IEF pattern of aralast.

The letter mentioned potential concerns which needed to be evaluated. These concerns included potential altered in vivo half life, altered biodistribution, increased antigenicity, and organ toxicity.

Then, on June 9 of this year, Baxter came to the FDA and presented results of a biochemical analysis of aralast, and I believe it is these data that Dr. Schwarz is going to talk about later.

These are the presentations that are going to follow. Dr. Marshall will present results from an expanded isoelectric focusing study of all three licensed products, as well as alpha-1 PI in plasma, and Dr. Brantly will discuss the detection of, and implications of, a modified alpha-1 PI in aralast and, as I just mentioned, Dr. Schwarz will present the results of Baxter's biochemical analysis of aralast.

Dr. Khoie from CBER will discuss results of a detailed consideration of safety reporting for all three licensed products.

Dr. Pierce will summarize proposed safety monitoring programs for al products, and post-marketing study commitments for two of the licensed products, and the rationale for the design of these studies.

Drs, Chang and Brorson will describe therapeutic proteins, both plasma derived and recombinant, that have known structural modifications and that have been licensed by FDA.

It is important to add here that, even though one product with a specific, well defined modification, is demonstrated to meet FDA criteria for approval -- that is, safety, efficacy, potency and purity, each product is considered on a case by case basis. In other words, a specific modification may allow approval of one product but not necessarily of another.

Even though we have yet to hear the presentations, let's keep in mind the following issues, since the committee will be asked to respond to these questions this afternoon.

Based on the differences in primary structure of alpha-1 PI, and the concentrations of alpha-1 PI polymers in these products, does the committee have any comments and/or recommendations regarding:

One, the adequacy of the requested or planned postmarketig commitment studies to evaluate the longer term safety and efficacy of alpha-1 PI products as measured by specified clinically meaningful end points;

Two, the adequacy of the proposed safety monitoring programs;

Three, any other suggested actions, for example, communications through labeling or other venues. Thank you.

DR. ALLEN: Thank you, Dr. Shrake, for the overview. Our next presentation is by Dr. Ewa Marszal, who will tell us about observations on marketed alpha-1-proteinase inhibitor products.

Agenda Item: Observations on Marketed alpha-1-Proteinase Inhibitor Products.

DR. MARSZAL: Good morning. I will be talking about differences among alpha-1 PI products, and the consistency of their manufacturing.

All alpha-1 PI licensed products are manufactured from human plasma, and that is why they contain contaminating plasma proteins.

In addition, they contain inactive alpha-1 PI species, such as polymers in latent form, and they have different kinds and different levels of modifications of the primary structure.

This slide shows protein composition of alpha-1 PI products. The analysis, denaturing, reducing polyacrylamic gel electrophoresis, staining with komase blue, and the intensity of the bands corresponds to the active amounts of the protein species.

Here I show analysis of three lots for each of the products, and the lots were randomly selected from 2003-2004 manufacturing.

Alpha-1 PI is the major component of all these products, and this analysis does not discriminate between the active and inactive forms.

The major impurity is albumen, and there are also minor other impurities. However, there is only one impurity -- IGA -- that we know that is causing serious adverse events. All these products have IGA warning on the label.

The specific activity of the products expressed in milligrams of active alpha-1 PI per milligram of total protein can be as low as .35 for prolastin, .55 for aralast, and .7 for zemaira, and this information is from the package insert.

This slide shows a molecular weight distribution of the species present in the products, and this analysis is size exclusion, high performance liquid chromatograph.

The area under the curve corresponds to the relative amounts of the species present in the product, and the elution profile is shown for one randomly selected lot for each product.

The major peak corresponds to the monomeric form of alpha-1 PI active and inactive, plus albumen. Albumen does not result from alpha-1 PI under these analysis conditions.

In zemaira, we can see the presence of alpha-1 PI dimer. in aralast and prolastin, we can see higher molecular weight species, which include polymers of alpha-1 PI and probably albumen, and also protein impurities contribute to this pattern.

Now, I would like to introduce isolectric focusing. It is an electrophoretic method. The proteins are applied to a gel and migrate through the gel in the electric field.

In isoelectric focusing, there is pH gradient. Alpha-1 PI is a negatively charged protein at physiological pH. It is applied in the cathode area, and it migrates toward the positively charged anode.

On the way, it is encountering an increasing concentration of protons, which neutralize its charge, and the protein stops at pH, where the next charge on the protein is a zero, and the protein focuses at this pH.

Alpha-1 PI, when analyzed by isoelectric focusing, shows high heterogeneity, and it is due to the presence of three polysaccharide chains, each of which can have two or three antennae.

Each antennae is capped with a sialic acid molecule, which carries a negative charge, as shown here in red.

So, different alpha-1 PI molecules with different numbers of branches have different charges, and they focus at different pH.

Heterogeneities, also caused by the presence or absence of N-terminal pentapeptide, which has two negative residues, and heterogeneities are also caused by natural alpha-1 PI polymorphism.

However, the effect of polymorphism is not seen in old plasma samples, and is not seen in the product. It is because dominating variants have the same charge.

These are dominant alpha-1 PI isoforms in plasma, M6 and M4 and M2. Although it is present in smaller amounts, it is also clearly visible on the gels.

All these forms have full length polypeptide chains, and the only difference is the number of sugar branches. M4 has one more negative charge converted to M6, and M2 has one additional negative charge compared to M4. So, they are present on the gels as three separate bands.

So, now let's look at isoelectric focusing analysis of plasma and the products. On the left-hand side you can see analysis of samples that were not manipulated in any way. On the right-hand side, you can see samples that were desyalyated(?).

In the plasma, we can see clearly isoforms M6 and M4, and also most likely M2, and the staining of this gel was with komase blue. So, above the bands, the staining shows the presence of other proteins that focus in this pH range.

The most similar pattern to that of alpha-1 PI in plasma is in prolastin, also two major bands, and they focus at the same pH as M6 and M4 isoforms in plasma.

In Zemaira, there are also two major bands, but they are shifted slightly toward down, which suggests certain modification.

In aralast, there are also two major bands. However, they focus at M4 and M2 positions. The simplest interpretation of such a result would be that actually in aralast, all alpha-1 PI is normal. However, the product is enriched in isoform M4 and M2.

However, a closer look at the gel suggests that actually these two bands correspond to isoform M6 and M4, creating an additional negative charge. So, they are shifted by one negative charge toward the anode.

In addition, we can see also the presence of normal M6 isoforms. This suggested that, in aralast, there are two forms of alpha-1 PI, normal and carrying an additional negative charge.

To test this hypothesis, we desyalyated the protein. In this case, we would expect to see two different kinds of polypeptide strains.

On the other hand, since this heterogeneity is only caused by the presence of the sialic acid after desyalyation, we should see only one band.

This is what we essentially see after the desyalyation. So, this is a western blot of plasma, and western blot was performed to be able to visualize alpha-1 PI, since it focuses in this region where there are many other proteins present.

On the other hand, for the products, we used komase blue stained gel. So, in plasma, there is a major band, but we also see minor other bands.

The band on the top most likely corresponds to polypeptides missing their N terminal peptide. However, there are also other faint bands here, which suggest that a small population of alpha-1 PI in plasma is also modified.

Now, the most similar pattern among the products is seen in the prolastin, and the major band focuses at the same pH as the major band in the plasma.

In Zemaira, the major band is shifted slightly toward down, which confirms that there is some modification of this protein.

In aralast, the major band is shifted by one negative charge toward the anode, suggesting that there is a modification.

Also, in aralast, we see a smaller amount of normal alpha-1 PI. In addition, we can see that, in zemaira and in prolastin, there is also a population of proteins carrying an additional negative charge, which suggests that there is modified protein present there.

However, the modifications here can be different, and there may be more than one modification, resulting in an extra negative charge, and the modifications in the product can be different from the modification that is observed in plasma.

Baxter carried out an extensive investigation to identify the cause of the atypical isoelectric focusing pattern of their product, and they will provide their data shortly.

One of the most important questions that we wanted to address when we learned about the atypical isoelectric focusing pattern of aralast was, if aralast tested in clinical trial had the same atypical pattern as it is showing right now.

To answer this question, we collaborated with Dr. Mark Brantly, from the University of Florida. He tested patients plasma samples from the clinical trial and provided us with the results. We were able to unblind the patients and analyze the data.

We found that, at the time of the clinical trial, aralast consistently was showing the atypical isoelectric focusing pattern that it is showing right now.

In addition, we analyzed a number of current lots for all the products, and historical lots for all the products. We found that each of the products shows the same isoelectric focusing pattern over time.

So, in summary, I would like to emphasize that alpha-1 PI products contain different levels of contaminating plasma proteins.

They contain different amounts of inactive alpha-1 PI species, such as dimers or polymers. They contain different levels of primary structure modifications.

Alpha-1 PI, in each product, differs somewhat from alpha-1 PI in plasma, or in terms of isoelectric focusing patterns, all because it contains polymers. Manufacturing of each alpha-1 PI product appears to be consistent over time, when judged by isoelectric focusing. Thank you.

DR. ALLEN: Thank you very much, Dr.Marszal. Questions for clarification? Okay, we will move on, then, to the next presentation by Dr. Mark Brantly of the alpha-1 Foundation. The topic is identification and possible implications of a human plasma purified anodal variant of alpha-1 antitrypsin.

Agenda Item: Identification of Possible Implications of a Human Plasma Purified Anodal Variant of Alpha-1-Antitrypsin.

DR. BRANTLY: Thank you very much. Let me just -- you will know more about alpha-1 A-trypsin by the end of this series of presentations, but I wanted to review just a little bit more about it.

It is a 394 amino globular protein. This is the protein structure. What is absent from here is the glycosilated forms, which we really don't know how they look on here.

This is the reactive site loop. It is under an enormous amount of stress. It sort of acts as bait for proteases, and the cleavage of this causes a major rearrangement.

It has a single cysteine in it, and it is the isoelectric point, for all of the variants fall within a pH range of 4.2 to 4.8.

The functions of alpha-1 A-trypsin is, one of the most traditional and primary ones that we focus on is the ability to inhibit a number of serine proteases, including neutrophil elastase.

In addition, recently it has been shown that it inhibits neutrophil defensins, which is a potent antimicrobial and a cytotoxic molecule.

It is now -- there is accumulating evidence that it has broad anti-inflammatory effects and, indeed, it may be one of nature's anti-inflammatory molecules. It inhibits lymphocyte trafficking. It appears, in vitro at least, to prevent LPS induced stimulation.

One of the names that it is also called is alpha-1 proteinase inhibitor, and that is important as we go through this talk.

Now, again, more isolate refocusing gels, because this is sort of the center of the story in this matter. One of the things I was noticing as we look at our gels a little bit differently -- so, I flip mine over a different way.

This is a schematic of what alpha-1 A-trypsin variants look like. Again, the IEF method is used for separating protein isoforms by charge. The isoelectric point is on the pH scale where a protein has no net charge.

Traditionally, it has been used for laboratory diagnosis of alpha-1 A-trypsin, but it is obviously very potent for understanding isoforms.

Now, in the first lane right here is M1M1, which is the most common variant seen in the human population. There are some different isoforms, which we will talk about a little bit more, but the major isoforms are the 4 band and the 6 band.

On the far end right here is the ZZ protein, which has a single amino acid substitution, which causes basically both a drop in the function and also the amount that is secreted.

You can see that an MZ individual right here actually has the combination of both of those isoforms in there.

So, an individual that is being augmented with alpha-1 A-trypsin would normally have an MZ pattern, and that sort of begins the story here.

This is actually what the isoforms look like. Again, there are basically five major isoforms. The two most prominent isoforms are actually the four and the six band, and they constitute approximately 85 percent of all the alpha-1 A-trypsin that is circulating. Again, the major differences are basically the carbohydrates.

For the seven and eight band, they are identical to the four and the six band, with the exception of a five amino terminal clip.

Now, let's just tell the story about how we identified the anodal drain in Aralast. Individuals with alpha-1 A-trypsin deficiency have commonly ZZ. So, the pattern on therapy, as it was expected, would be MZ.

The E region is actually -- we have a known variant for E, which is a different amino acid substitution. This E region variant was identified in two different laboratories, my lab, and a lab at Temple University.

At first, when I saw this, I thought I was looking at a genetic modification. So, I actually did genotyping and found that the subjects were ZZ, and then surmised that it was probably some sort of post-translational modification. As it turns out, after investigating, we found that all the subjects were on Aralast.

This brought some concerns, in that originally I was very involved in the Alpha Therapeutic Corporation's development of their API, which subsequently was licensed to Aralast.

Originally, when I did the analysis of that particular product, it had an MM pattern to it. So, there was a change in the pattern from the time that we did initial analysis of the material until now.

So, one of the big questions that the FDA had was, is this the same material that is used in the pivotal trial, because obviously that would be a big problem.

So, let me tell you a little bit about the Alpha Therapeutics Corporation pivotal trial. Again, as I mentioned, this ended up being the same material that Aralast is.

The clinical trial was run from 1996 to 1999. There were 28 deficient individuals that were enrolled for a six-month study. There were two early drop outs.

The comparitor was Prolastin. The primary outcome variable was not compared, that is, the total and the functional amount of atrypsin.

That was the central laboratory for this particular study, both at the NIH, and subsequently when I moved to the University of Florida.

It was a two-arm cross over study, where the ATC API had 13 individuals, and the Prolastin had 13 individuals. This period right here was 12 weeks and they were crossed over to ATC API at the end of this thing, and again, it was a 24-week study.

We actually had an opportunity to analyze all the samples that we had available. We had 11 of the 13 samples. I wanted to select just a specific time prior to cross over, because I think this is illustrative as far as looking at this.

This is a lot of Aralast, where you can see the major bands here. Here is an individual with some prolastin, and here is an individual that is on Aralast, with very little of the M bands sort of being visually picked up. You can actually very easily -- you don't have to have the key to figure out which patients are on prolastin and which ones are on Aralast. Again, here are the Z bands running down here and a standard.

We used these bands, actually -- I did then some metric studies on the banding patterns, and looked at the proportion of the E as compared to the N, and the results are here on the right-hand side.

A densitometry was done at the six week point, and I am just showing the data on the ATC API patients, rather than the prolastin patients.

What we found was that approximately 76 percent of the total nadir alpha-1 A-trypsin -- this was done right before the next dose -- was the modified form, that you can see right here.

This was consistent with our observations, and also, importantly, consistent with what we anticipated that Aralast had, which the vast majority of the material was the modified form.

Now, the pivotal results, the ATC pivotal result was in a publication, and I just want to go over them briefly.

It demonstrated that the antigenic and the functional amount of alpha-1 A-trypsin, based on the antiprotease activity, was similar to comparator prolastin.

There were no serious safety signals in this small, short study and, following six weeks of the ATC API, the modified material was about 76 percent of the materials total.

The half lives were similar between ATC and API. There was a statistical difference between them, suggesting that there was a longer half life for the ATC API, but the confidence bands definitely overlapped each other.

Now, Baxter did a great deal of elegant work in trying to figure out the cause of this anodal variant. As it turns out, the anodal variance is secondary to the loss of the C terminal positively charged lysine, which is located here in the structure. That loss is presumably secondary to carboxy peptidase activity, which is a contaminant.

Now, the question is whether the loss of this terminal lysine is important or not, and I think that is sort of the central issue of all the presentations here.

So, let me show you some structural studies that sort of look at what the lysine 394 is before and after protease cleavage, because that is sort of a key component.

One of the things you can see is, even after cleavage and the significant rearrangement of the molecule, the 394 amino acid does not move. It basically is located -- this is two different crystal structure salts. So, we know this actually is probably a fixed position rather than a non-fixed position, in that a non-fixed position, actually you can't get x-ray coordinates on, because it is moving around too much.

The other thing is that it appears to be in range of the serine 292 as far as hydrogen bonding. I didn't bring some cartoons to show you rotation, because I thought it was a little bit too much to go over here, but it is quite important.

One of the things I want to focus in on is this molecule right here, which is the cleavage product, which is also called the C36 peptide.

There are now a number of studies that are coming out suggesting that this is a biologically active molecule that may actually be a weak agonist on the TLR-4 receptor, suggesting that it may actually block LPS mediated events.

So, we don't know whether this may cause it to lose more of this peptide, in other words, to be freed up, or whether it changes its biology.

One of the other things that I had the opportunity, because my lab does a lot of this, and we do a lot of work for the industry is, I have had the opportunity to basically test a number of products, even some of the products that the FDA doesn't have access to yet.

This is an isoelective focusing gel looking at five different forms of alpha-1 antitrypsin. There is Acommida(?), which is an Israeli company, that is developing alpha-1 A-trypsin, and Prolastin, Zemaira, Aralast, and Trypsone, which is made by Griffels(?).

On the far sides here are M-1 standards. In other words, this is plasma from a normal individual. You can see, here are the major bands. That is the 2 band right here, the 4 band, the 6 band, the 7 band and the 8 band, that you see.

You can see that the Acommida material has a similar band dependence, but has this material right here, which is not seen in the plasma.

You can also see that Zemaira is also. So, this group of three are very similar. If you look at both Aralast and Trypsone, they have a similar banding pattern, where a majority of the alpha-1 A-trypsin is actually this anodal variant here.

These are high resolution scans. So, we can actually see a little bit better this distinct difference, and that these are definitely different.

So, again, none of these products are exactly the same as native alpha-1 A-trypsin, but more than 65 percent of the trypsin in Aralast is the modified form, presumably.

We also know, from studies, that both prolastin and zemaira have some of the modified forms. It is in the range of two to six percent, and there are other known modifications in these products.

So, is there reason for concern using a modified form of alpha-1 A-trypsin for augmentation therapy? number one, i would like to remind you that this study -- in fact, all the studies involving this, were very small clinical studies that were done to establish the safety profile.

The modified form is 30 percent more in aralast than it is in prolastin, and I think that is an important thing to remember.

There is the potential of antigenicity because of the loss of the surface amino acid, and we know that is for sure.

Now, alpha-1 A-trypsin, there are other variants that don't cause antigenicity. So, i think that we are comforted by the fact that other molecules may not cause antigenicity.

The charge difference may make a significant difference, though, in sort of the biodistribution of the molecule.

Obviously, a highly positive charge may be really cleared. .We just don't know at this point. Again, this dimension, the cleavage of the modified form creates a modified C terminus, which may end up being important.

One of the things we clearly know is that alpha-1 A-trypsin has many functions that we haven't quite figured out. I would also remind you that it is the second-most abundant plasma protein.

There is reason to believe -- and unfortunately there is reason to believe that ATC did their animal studies, their animal tox studies, probably using the other form of alpha-1 A-trypsin, not the form that went into the clinicals.

So, in summary, the anodal variant is a modified form of alpha-1 A-trypsin. It is a truncated C terminus, which occurs probably during purification.

At least 65 percent of the aralast is the modified form. Then, truncation could potentially modify antigenicity tissue distribution, clearance and stability.

This modified form of alpha-1 A-trypsin may or may not have all the functional properties of native alpha-1 A-trypsin.

I think that careful monitoring of patients taking this group of products is really warranted. Thank you very much.

DR. ALLEN: Thank you, Dr. Brantly. Questions of clarification of Dr. Brantly on his presentation?

DR. EPSTEIN: One question, Mark. You pointed out that the Griffolds product has essentially the same protein modifications as the aralast. Are there any safety data in Europe on the Griffolds product?

DR. BRANTLY: I haven't had access to it. So, it is just unknown. The commonality between -- and I think the Baxter folks can tell you more -- they are basically essentially the same product, and they had the same protocol and that kind of stuff. So, we don't know.

They have a much smaller core and, if I recall, all the Griffolds product is being used in either South America or in Spain.

DR. DI BISCEGLIE: My question is, what is the significance of, or how important, is measurement of the activity in trying to assess differences or clinical utility, or comparisons between products?

DR. BRANTLY: Actually, we have taken a molecule that is a highly conserved molecule, and we picked out, in our puny human terms, picked out one function that we think is the critical function in this particular molecule, which is an antiprotease activity.

Do I think it is important? I think it is important. Do I think that it is the whole story? Absolutely not.

I think that we have chosen this one function -- and indeed, a very important function -- to sort of focus in on.

Quite frankly, it is very difficult to measure many of the other properties that this molecule may have at the present time.

DR. DI BISCEGLIE: In the improved indication, we are trying to prevent emphysema. Is this activity not thought to be the pivotal activity in causing that disease? Yes, the molecule might have other activities, but in this disease, isn't that the most important activity?

DR. BRANTLY: That is the hypothesis. I think that, as a scientist, I have to say that we may be wrong. Again, this is based on a hypothesis of the anti-protease homeostasis hypothesis, and that is that it is inhibition of neutrophil elastase that causes the problem.

We know that these molecules have potent anti-inflammatory effects. So, it may be a chicken and the egg process also, in that the recruitment of neutrophils into the lung may be a major key step in going through this.

So, the anti-inflammatory properties may be more important. It is just a matter of how you look at the molecule and its major process. I have been humbled enough time in my hypothesis that I am a little more circumspect now.

DR. LAAL: Does the structural change lead to an immune response that would eventually alter the function of your molecule?

DR. BRANTLY: We know, from the ATC study, that the amount of alpha-1 A-trypsin is similar to what it was with prolastin.

That suggests that there are not neutralizing antibodies. So, its function specifically for its anti-protease effect, and actually the amount of it, is not altered, suggesting that at least in that study, that there was antibody formation.

I did do a crude antibody test during this study, but unfortunately, one of the things is there is not -- you can't get a positive control for this. So, I don't know how specific it was and how sensitive the assay was for antibodies. I feel pretty strongly, from the data, that there is no evidence of neutralizing antibodies.

DR. SZYMANSKI: I might have missed in your presentation, but how many polymers did you find in there?

DR. BRANTLY: You know, isoelective focusing is not very good for polymers, because they don't tend to migrate out into the gel.

Better ways to look at polymers, actually, are other methodologies, either HPLC or actually non-denaturing gels.

DR. SZYMANSKI: Would the polymers be considered dangerous?

DR. BRANTLY: It is not known. There is a body of literature to suggest that polymers are pro-inflammatory. There are some suggestions from other parts of the literature that suggest they may not be pro-inflammatory. .So, I think that the process -- I mean, I think the jury is a little bit out on that aspect of the polymers.

One thing I would remind you is that in general our RE system hates aggregates and polymers. So, usually they are cleared by first pass. I don't believe anybody has been able to detect increases of polymers in people who have been augmented.

I also want to remind you that polymers naturally form, actually, in normal sera, too. Particularly, you can get polymer formation for Z individuals that occur.

DR. DOPPELT: I think you said that, in the initial toxicity studies, that the product was different from this current product. Is there something different in the manufacturing?

DR. BRANTLY: This is the part that I can't stitch together for you because there is sort of a black box. I don't believe that Baxter can stitch that together either.

In 1995, when ATC came to me to help them with the analysis of their drug, I did isoelectric focusing on that drug, and actually did denaturing gradient gels, and did an extensive characterization of that particular molecule. It was very similar to prolastin at that point.

Between 1995 and 1996, my understanding is a number of manufacturing changes were made, and I did not analyze -- I was not asked to analyze those samples, and went forward with the clinical trial at that point, not realizing that there had probably been a change in the molecule.

From the timing, I would guess that the molecule used for the animal studies was not the same as it was for the clinical study. Actually, when I was given the material, the animal studies had already been done.

DR. SZYMANSKI: I would just like to know how much is this alpha-1 antitrypsin concentrated in the products versus the normal plasma.

DR. BRANTLY: Many, many, many-fold. Example, it depends on the manufacturer, quite frankly, but usually there -- it is re-suspended and then fused in a range from, let's see, 50 milligrams per ml, and normally in the human sera it is about 150 milligrams per deciliter. So, it is substantially concentrated. The typical dose is about 4.5 grams for a 60 kilogram individual.

DR. ALLEN: Thank you very much. We will move on to our next speaker at this point, Dr. Hans Peter Schwarz from Baxter Health Care, characterization of Aralast compared to other alpha-1 proteinase inhibitor preparations.

Agenda Item: Characterization of Aralast Compared to Other AIPI Preparations.

DR. SCHWARZ: Good morning, members and guests. I would like to address, in the next 15 minutes or so, modifications to primary structure on alpha-1 Antitrypsin in all commercial products, the potential cause of the micro heterogeneity of alpha-1 A-trypsin in Aralast, and I will close with an elucidation on the potential implications of the microheterogeneity found on alpha-1 antitrypsin molecule in Aralast.

Now, this is a two dimensional IF gel, what I would call the united colors of alpha-1 antitrypsin. Here, aralast and prolastin were labeled with a specific dye, which enters co-linked bounding with epsilon amino group of lysine, and you nicely see aralast in green, and prolastin in red, and the slight anodal shift of all the isoforms. At the same time, this technique allows you to identify different contaminating proteins.

These 2D IF gels confirm what you saw presented by the previous speakers, that there is an anodal shift of the alpha-1 PI molecule in aralast. This is just a magnification of the mobility of the alpha-1 antitrypsin protein in aralast toward the anode, compared to prolastin in red.

Now, what are the modifications which have a potential impact on the protein charge, secondary structure modifications based on the glycosylation differences, and primary structure modifications such as deamidation, cysteine modification, and removal of terminally located charged amino acids.

This is a summary of data comparing the N glycan structures on the three commercially available products, aralast, prolastin and zemaira, by an HPLC analysis, and you will appreciate that there is no difference between the products. This is also confirmed by mydetov(?) analysis.

So, there is no difference in N-glycan profile between aralast, prolastin and zemaira, and the pattern is similar to that found for antitrypsin in plasma also, although I am not showing you the data..

So, the first message I would like to convey to you is that the N glycan pattern of all three antitrypsin products used for the treatment of hereditary emphysema is similar to that of plasma and, therefore, ethanol fractionation and downstream purification has no impact on any glycan differences.

Now, this is a high resolution IF gel, and it shows you two things. Number one, the consistency of the IF pattern of aralast over the years.

This is a lot made in 1996, and then this is a lot of 2001, 2004, 2003 and 2002. Second, the purpose of this slide is to correlate the N glycan pattern to IF bands.

Prolastin in plasma, N6, as we heard just now before, has three B antennaric glycan structures attached and, in four, two B antennary and one 3-antennary structure. This is both in prolastin and aralast, and we actually did analysis of this gel by cutting out the gel and subjecting them to MS.

Now, what is the situation in the banding pattern of aralast? This band, which co-migrates at the level of M-4 in plasma and prolastin, consists or has the three B antennary glycan structures attached, similar and identical to M6, while this band, in aralast, has the two B-antennary and one 3-antennary structure attached.

So, this band corresponds to M6, present in plasma and prolastin, and this band corresponds to M4. Then there is a shift toward the anodes, and the N-glycans are therefore not responsible for the IF pattern characteristics of aralast.

Therefore, we started the investigations to look into potential changes on the protein backbone. First, deamidation.

Deamidation potentially results in an additional negative charge. it is the non-enzymatic conversion of asparagine 116, or asparagine 314 to aspartic or isoaspartic acid.

This slide shows the data for aralast, zemaira and prolastin, both by a quantitative assessment, which was then confirmed by MS analysis.

In aralast, six percent of the molecule is deamidated to one isoaspartate, in zemaira seven percent and, in prolastin, 10 percent of the molecules are deamidated.

So, deamidation is one of the primary sequence modifications which are present in all commercial products. What about cysteine modification? This is the attachment of another molecule, 0-cysteine to cysteine 232.

We came across this because here we did another high resolution IF gel, and using zemaira, prolastin and aralast under reducing and non-reducing conditions.

You will see that, in prolastin, under reducing conditions, there is a slightly anodal shift, which suggests that the cysteine is too stimulated in prolastin and in aralast.

We confirmed this by MS analysis. Aralast and prolastin have another cysteine attached to the cysteine 232, while this is absence in zemaira, and this is probably due to the purification process of zemaira, which includes denaturing agents.

So, aralast and prolastin, as well as alpha-1 antitrypsin in plasma, exhibit cysteinylation at 232, and this is not detected in zemaira.

What about the C terminal truncation, which would result in the loss of a positively charged lysine, and the des-lys isoform of antitrypsin, which we have identified.

Here is, again, an electrospray analysis showing the lack of the C terminal lys in aralast, as well as in prolastin, as well as in zemaira, and here are some peptides with the presence of the lysine. Using an HPLA analysis, you can easily quantify the lack of lys on these products.

Our analysis resulted in the following conclusion: The des-lys antitrypsin is present in aralast at the level of six to seven percent, in zemaira around six percent, and in prolastin about two percent.

Now, this summarizes the primary and secondary structural changes we have identified on the alpha-1 antitrypsin molecule in all the commercial products.

There is no difference in the glycoisoforms when you compare aralast, prolastin and zemaira in plasma. However, I would like to point out that, in the course of this investigation, we identified, in plasma, novel site specific N glycan patterns.

For example, for the first time, tetra antennary structures were attached to antitrypsin, as well as Lewis X structures in plasma alpha-1 antitrypsin.

Deamidation is present in alpha-1 antitrypsin in all the products available. It is not possible to assess this in plasma.

We are currently working on identifying the lack of the C terminal lys in normal plasma, but I will show you later data which suggests that in other tissues, in other body fluids, such as BL, actually the C terminal lys is removed.

The cysteine, decysteinylation at position 233 is present in aralast and prolastin, absent in zemaira and present in plasma.

So, all antitrypsin molecules in the different commercially available products show some differences as compared to alpha-1 antitrypsin from plasma.

What about the potential cause of this particular micro heterogeneity, especially expressed in aralast? There is a group of enzymes, a family of enzymes, the so-called carboxy peptidases, which cleave C terminal basic amino acids, such as lysine and arginine from all kinds of proteins and peptides involved in hormone maturation and complement activation.

I just want to mention three carboxy peptidase M. This is a membrane bound carboxy peptidase, which is highly expressed in lung tissue, but also present in the BL fluid. Carboxy peptidase N, which is active in plasma, and found at the concentration of 30 micrograms per ml, and carboxy peptidase U, which is also known as TAFI.

Now, this slide shows and outlines the presence and concentrations of two main carboxypeptidases in plasma, as well as in different fractions of the fractionation, plasma starting material and then, in blue, carboxy peptidase and, in red, TAFI.

I just would like to point to fraction IV-1, which is used for the manufacturing of alpha-1 antitrypsin. So, relatively high concentrations.

I would like also to mention that carboxy peptidases are below the limit of detection in final containers of all products available.

Now, this slide shows the ethanol dependence of the C terminal lys cleavage induced by CPN. Here we used a specific assay which can measure the lys removed by the action of CPN, and increasing amounts of ethanol will facilitate and accelerate removal of the lysine, in pink, in prolastin, and in blue, aralast. So, there is more lysine removed from prolastin because it has more to be released.

Now, this is an important slide because it shows actually that CPN can remove the C terminal lysine also in absence of ethanol, but the reaction is highly stimulated in the presence of ethanol.

Interestingly, recombinant yeast derived alpha-1 antitrypsin seems to be more susceptible to the action of CPN.

This is a time course of the removal of the lysine from recombinant alpha-1 antitrypsin. So, in five minutes, more than half of the lysine is removed.

This high resolution IEF shows the generation of anodal isoforms of alpha-1 antitrypsin by the treatment of recombinant membrane bound CPN.

This lane is aralast, and in lane zero is an experimental lot similar to prolastin, and increasing concentrations of carboxy peptidase M applied.

So, you see a shift of the M6 band, you see a shift of the M4 band, interestingly also a shift of M7 to M7* and a shift from M8 to M8*. So, carboxy peptidase M is very potent in generating these specific isoforms which lack C terminal lysine.

Here is a quantitative assessment of the conversion of M4 to M2* and M6 to M4*, upon the action of carboxy peptidase M.

Now, this is very new data, and you have not this in the package. This is alpha-1 antitrypsin isoform peptin obtained from human PL samples.

Here a highly sensitive IF gel was used to detect alpha-1 antitrypsin in bowel samples from subjects not on replacement therapy, and alpha-1 antitrypsin was detected in all bowel samples.

Interestingly, the IF pattern resembles the alpha-1 antitrypsin shift observed in aralast here in lane two, and here are the patient samples.

This suggests that an isoform shift with the removal of C terminal lysine may occur naturally, and is possibly induced by the action, or the presence, of membrane bound carboxy peptidase M.

So, in summary, all basic carboxy peptidases can cleave to C terminal lysine from alpha-1 antitrypsin. The cleavage of C terminal lysine occurs also in absence of ethanol.

Ethanol enhances the reaction of CPN, approximately 24, and if it is 15 percent ethanol, there is a 20-fold effect on both KCAT and KM in the presence of 15 percent ethanol.

CPN is the most likely candidate causing the C terminal lyse cleavage in plasma, and I have no time to give you the reason for this statement.

I would like to refer also to an interesting paper published in the JBC, 1962, where they describe effects of ethanol both on KCAT and KM, regarding the activity of carboxy peptidase B.

Now, potential implications of the micro heterogeneity of alpha-1 antitrypsin observed in aralast, we did extensive preclinical in vivo and in vitro investigations to rule out concerns regarding the additional extra negative charge.

This is just a summary of a very extensive bioinformatic investigation. Actually, the investigations were performed in two different institutions, and they came to the same results. Let me walk through this.

There is no conservation between species of the C terminal lys in alpha-1 antitrypsin. Therefore, the lys 394 is unlikely to play a major structural or functional role.

3-D analysis of available structures for alpha-1 antitrypsin, and its complex with protease do not support a major structural role of lys 394.

The C terminal loop region is stabilized by a hydrogen bond network, in which lys 394 is not involved. So, the conclusion is that lys 394 plays no major structural role in alpha-1 antitrypsin, when uncleaved and cleaved, as well as when forming complexes with its proteases.

This slide summarizes the functional relevance of the C terminal truncation. It shows the alpha-1 antitrypsin dependent inhibition of porcine elastase, so the inhibition of the porcine elastase as a function of the concentration of alpha-1 antitrypsin.

You will appreciate that, for prolastin, zemaira, aralast and an experimental lot we performed with it, that has 100 percent des-lys truncation, the curves are superimposable. So, lys, terminal lys, has no effect in any way on the capacity of alpha-1 antitrypsin to neutralize porcine elastase.

Now, thus, the high degree of C terminal truncation, and introduction of an additional negative charge has an impact on metabolic clearance in rats. There is comparison of different alpha-1 preparations which had a different degree of C terminal lys truncations.

Distribution of the lung, we did rat BL studies. There was a clear comparability of alpha-1 preparations with different degrees of C terminal lys truncation, including the 100 percent des-lys alpha-1 PI in rat BAL studies.

Diffusion from vasculature into the interstitium, it is an interesting model which is called the skin blister suction model, comparability of alpha-1 antitrypsin preparation with different degrees of C terminal lys truncation, including the 100 percent des-lys form in guinea pig suction blister models.

Finally, we did confocal studies to assess the diffusion of these different forms of alpha-1 antitrypsin from the vasculature into the lung tissue, and also here, there was no difference observed.

So, if I may include, alpha-1 antitrypsin in all products approved for augmentation therapy demonstrate at least one primary structure modification.

The des-lys alpha-1 antitrypsin is induced by the action of carboxy peptidases, and the ubiquitous presence of these enzymes in plasma, as well as in lung tissue will likely result in exposure and, hence, tolerance to the des-lys alpha-1 antitrypsin form.

The des-lys is one of many known isoforms of alpha-1 antitrypsin that does not affect inhibitory activity, immunogenicity or essential functions of alpha-1 antitrypsin. Thank you very much for your attention.

DR. ALLEN: Thank you, Dr. Schwarz. We will entertain clarification questions.

DR. DI MICHELE: Thank you, Dr. Schwarz. As usual, it is beautiful, elegant biochemistry, and it takes me a while to catch up, but I actually have a couple of clarifying questions.

Based on what you presented, and the fact that the same isoform sort of shift occurs naturally in alpha-1 patients, antitrypsin deficiency patients, does it appear to be an inherent molar imbalance between the carboxy peptidase and the amount of protein substrate that you think might be causing the shift?

In one case, in your manufacturing, you seemed to be enhancing the carobxy peptidase end, and in the other case, naturally occurring, there seems to be just naturally less protein. Is that what might be going on here?

DR. SCHWARZ: That is a very good question, like always. It might be a combination of different effects. As I demonstrated, there is a clear effect of ethanol by decreasing TTN and increase TKCAT.

On the other hand, the removal of the C terminal lys is time dependent. So, the obvious question is, you know, how long is the alpha-1 antitrypsin in the fractionation intermediates, exposed in the presence of ethanol to the carobxy peptidases.

Obviously, manufacturing procedures and ethanol are almost generic, but there are slight differences in ethanol fractionation between the manufacturers.

Regarding the naturally occurring isoform or truncation of lysine, you know, it might well be that there is a conformational change when the alpha-1 antitrypsin molecule travels from the vasculature through the lung tissue into the alveolar space, which allows exposure of the C terminal lys in a way that makes it more accessible to the action of membrane bound carboxy peptidase, but this is speculative.

DR. DI MICHELE: Thank you. The other question I have is, in addition to there being this anodal shift, there appears to be, of course, also a shift in the relative concentrations of the M6 and the M2 isoforms, and a shift toward an increase in the M2, a decrease in the M6.

You showed a lot of physical chemical properties that haven't been altered by that shift, but the question is functionally, is there any reason to believe that an M2 isoform might have increased immunogenicity when present in higher concentrations, let's say, relative to the native molecule.

Secondly, have you done any studies on the sort of inhibitory activity on, for instance, neutrophil elastase, and the relative effect on that functional activity from a change in isomer in the product.

DR. SCHWARZ: I will answer your second question first. We used the porcine elastase assay, and here we compared, as you remember, different products, as well as the one that lacks totally des-lys. So, here there is no difference, but we have not done neutrophil elastase.

Regarding the immunogenicity of specific isoforms -- if this is your question -- if at M2 it would be more immunogenic than at M6, I cannot answer this.

All that I can say is that we are not satisfied, as you know. What we did were classic antigenicity studies, where we immunized rabbits with prolastin or aralast and then did both IF gels as well as page, and both antibodies recognized the same banding pattern. That partially answers your question, because we did IF also. So, they recognized the --

DR. DI MICHELE: You did some animal immunogenicity studies. The only question, though is that, I know from some of the materials that we were given, that fraction -- you didn't initially use fraction four as your starting material, and I was just wondering why that change was made.

DR. SCHWARZ: The starting material, prolastin is made of fraction IV-I, and aralast is made of a combination of IV-I plus IV-IV. That is the difference.

DR. KLEIN: Just so I am clear, the animal immunogenicity studies were done with the material that was used in the phase III trial?

DR. SCHWARZ: It is retail material, it is commercially available material. If I may just clarify something, these gels are extremely high resolution.

So, the charge difference, from a biological relevance, is negligible. While one sialic acid gives a PI change of .06, the lys truncation PI difference is .05.

DR. ALLEN: Thank you, Dr. Schwarz. We will move on, safety reporting for alpha-1 proteinase inhibitor products, Dr. Tina Khoie, Food and Drug Administration.

Agenda Item: Safety Reporting for Alpha-1-FI Products.

MS. KHOIE: Good morning. My name is Tina Khoie, and I am a fellow within the office of biostatistics and epidemiology within CBER. My presentation today will focus on the alpha-1 proteinase inhibitor product safety reporting.

Before I begin, I would just like to acknowledge my colleagues in the OBE office, Catherine O'Connell, Craig Zinderman and Robert Wise, who were all invaluable resources as I was preparing this presentation.

My primary objective for today is to present and summarize and describe all of the spontaneously reported adverse events from alpha-1 proteinase inhibitor post-licensure safety surveillance.

To start off with, in the next three slides, I will be giving some definitions of some important terms that I will be referring to.

I will also review the information in the alpha-1 proteinase inhibitor safety labeling, and I will review the error surveillance system.

First, I would like to make sure everyone is familiar with some of these important terms. First, an adverse event is any unwanted outcome which is associated with an alpha-1 proteinase inhibitor product, irrespective of causality.

A report is one adverse experience, and it excludes duplicate and follow up reports.

A serious adverse event is one that results in death, a life threatening experience, hospitalization or prolongation of hospitalization, disability or incapacity, a congenital or birth defect, or other important medical events, based on clinical judgement.

Finally, an unlabeled event is one that is ont listed in the professional package insert.

So, what adverse events are currently listed in the package inserts of the alpha-1 proteinase inhibitor products?

This table provides an overview and, therefore, does not list all the adverse events in each of the package inserts.

Although there are differences in the contents of the package inserts, I will review seven of the adverse events which are listed in each of the product package inserts for prolastin, aralast and zemaira.

They include allergic-like reactions, chest pain, dyspnea, or shortness of breath, dizziness, fever, rash or pruritus, and viral transmission.

I am going to give a brief overview of AERS, the adverse event reporting system, for those who may not be familiar with it.

AERS is the FDA's passive, post-marketing safety surveillance system, and by passive I mean that, for the most part, reports are submitted on a voluntary basis.

The AERS data base contains adverse event reports for all drugs and therapeutic biologics dating back to 1969. The reports that it contains all direct reports, which originate from consumers and health care professionals, all manufacturer reports through October 1997, and all serious manufacturer reports since November 1997.

Since November 1997, non-serious reports submitted by manufacturers in period reports have no longer been entered into the AERS data base. However, we can retrieve those reports by going back to the original submissions.

So, safety reporting in terms of our methodology was obtained by querying the AERS data base for all adverse event reports in which an alpha-1 proteinase inhibitor was listed as a suspect product. No other restrictions were place in this query.

As I mentioned previously, ARES no longer contains the non-serious adverse event reports submitted in periodic submissions since November 1997.

Although prolastin has been on the market for a significantly longer period of time, I was able to retrieve the non-serious periodic reports for a time period in which all three products were licensed, and that was between January 2003 and July 2005.

The data that I retrieved from the reports included demographics, adverse reaction MEDRA PTs, and lot numbers. MEDRA PTs stand for medical dictionary for regulatory activities preferred terms, and these are an internationally recognized set of medical terms that were instituted in November 1997, and really helped to form a standardized way of classifying and communicating about medical information throughout the medical product regulatory cycle.

The results of the AERS query and the periodic report search resulted in the identification of 806 prolastin reports since prolastin's approval in 1987, 38 aralast reports since its approval in December 2002, and six zemaira reports since its approval in July 2003.

Ninety percent of all the alpha-1 proteinase inhibitor reports are domestic, 10 percent foreign, and prolastin was the only product which had foreign reports.

From this point on, I will be focusing primarily on the domestic reports. In terms of demographics, 12 percent of domestic reports were classified as serious, 67 percent as non-serious, and 21 percent of reports had this information missing.

Fifty-one percent of all reports involved patients between 30 and 69 years of age, one percent of reports involved patients less than 30, and one percent greater than 70 years of age, with 47 percent of reports having age data missing.

Forty percent of reports involved females, 35 percent males, 25 percent of reports had missing data for gender.

So, this table shows the actual number of domestic reports that were received by the FDA between January 2003 and July 2005, the time period in which all three products, at some point, were licensed during these three years.

The reports are broken down in terms of the ones that were serious and non-serious. I would just like to point out that the date that we went by was the date on which the FDA received the reports.

A large number of these reports came in as periodic submissions, and prolastin's periodic submissions are submitted on an annual basis, and aralast and zemaira at this point are submitted on a quarterly basis.

So, the data from 2005, the year 2005, includes the periodic reports that we received, the annual periodic report that we received from prolastin already for 2005, and about three quarters of the periodic reports for aralast and zemaira, which were received also in 2005.

Although these three products are listed on the same table, I would just like to stress that we must not compare the numbers between these three products.

These numbers have not been adjusted for differences in market share, and so we can take a look to see how many reports we are dealing with, but we should not be comparing numbers between the three products.

Now, we will review what adverse events were reported. The first two tables will review the most commonly reported adverse events, and some of the more medically important adverse events but, again, will not represent all adverse event MEDRA PTs that were listed in all of the adverse event reports, and these are domestic reports.

There were over 250 different MEDRA PTs listed amongst all of the domestic alpha-1 proteinase inhibitor adverse event reports. This table just lists the 10 most commonly identified MEDRA PTs since product approval.

The number of reports that you see represent the number of reports that identified each of these MEDRA PTs. Keep in mind that reports generally listed more than one MEDRA PT.

The most common adverse MEDRA PTs included, from this list, dyspnea, headache, and fever. Once again, we must not compare the numbers between these three products, because we have not made adjustments for differences in market share, nor in the length of time that the product was on the market.

This table, rather than listing the most common adverse events, lists some of the events which were judged medically to be more important, based on the level of severity commonly associated with these types of events.

It is important, once again, to note that we must not compare the numbers between these three products, for the fact that we did not take into account the differences in market share, nor in the differences in time on the market.

Finally, I would just like to conclude by emphasizing some of the limitations of passive surveillance and spontaneous reporting, some of which were nicely reviewed by Dr. Gaines.

First, there is significant under-reporting, most likely greatly underestimating the actual incidence of these events in patients who were on these products.

Secondly, there are reporting biases involved. Newer products, that are newly licensed, there is generally a greater enthusiasm for reporting for these products.

In addition, if there are clinical studies underway, AERS tends to capture more of these reports, since clinical study adverse events are required to be reported.

Finally, reporting biases involved the fact that reports that are sent into the FDA may differ in some characteristics compared to those that are not sent in to us, perhaps in severity of the event.

Another limitation is the fact that reported events are unverified and often they are incomplete. For example, we received several chest pain reports, but many of the reports lack some important background clinical information, such as if EKGs were done, lab tests, and what those results were, information on the temporal onset of the event, physical exam result, past medical history, concurrent medications, and so forth.

So, without that kind of important background, it is hard to determine the type of chest pain that was involved in the report.

Another limitation involves the fact that reported events may be related to confounding factors such as underlying medical conditions and concurrent medications.

Finally, causal inferences are not usually possible when using passive surveillance data, because a direct link between the event and the product is not always possible.

So, the consequence of all of these limitations, the bottom line is really that numerical differences amongst these three products require very cautious interpretation.

I would just like to thank my colleagues within OBEE and ORRR, and I will take any questions or comments. Thank you.

DR. ALLEN: Thank you, Dr. Khoie, and I think you have summarized very nicely the problems with passive surveillance.

Clarification questions only, please? The next presentation is post-marketing study commitments. I would assume that we will get into some of this other.

DR. QUINN: Two questions, very briefly. One, why do all the package inserts say that there is an increased risk of viral transmission with this product, yet it never came up in any of the adverse event reporting. That is part one.

Part two is, if you looked at it historically, even with passive reporting -- you did the last three years, but if you went back further, was there a shift or an increase in reporting? Is that because passive reporting is getting better or is it because the product is actually having increased side effects?

DR. KHOIE: Regarding your first question, the viral transmission being on each of the product package inserts was related to the section of the label that provides caution and warnings of potential transmissions, and not necessarily that they had actual cases of viral transmission from pre-licensure studies.

DR. QUINN: Do you know why?

DR. KHOIE: I would have to defer to OBRR.

MS. STEFANO: Basically, there are very often times when, to exercise -- to err on the side of caution, we will have class labeling statements.

In other words, if you are a member within this class and you are using human based products, or animal based products, depending, there is this class labeling.

DR. KHOIE: The second question, the trend in the number of reports prior to 2003, I do have the numbers of the adverse events combined for each of the products.

I did take a look at the number of events for each of the products individually. Prolastin, having been on the market for a substantially longer period of time, I can actually show you that graph. I have a reserve slide.

This actually just shows the distribution of all the adverse event reports for all the products. It breaks it down into foreign and domestic, but the majority of these are prolastin, especially prior to 2003, and we see that -- I just also have to say that these do not include the periodic reports that I had mentioned I had retrieved.

More recently, since the late 1990s, there has been no significant increase. We see a blip in 1995, and that can be partly explained because we know there was an ongoing clinical trial at that time. So, errors would have been captured in more reports, because those reports were required to be submitted.

DR. KLEIN: I was just wondering if there are additives, stabilizers, incipients, that are different in the three products that you described that might account for some of the adverse events that are reported, like headaches and dizziness.

DR. KHOIE: In terms of the additives, I would have to defer that question to someone from OBRR.

DR. SHRAKE: With regard to the excipients, there certainly is some difference, but there is no reason to believe that the excipients are the source of the adverse events, actually.

DR. DI MICHELE: I was just wondering, I can imagine that some of these adverse events are sort of contaminated by the IGA contamination and some of them may be IGA related.

I suspect that, if you don't have sex, age and seriousness of event data, there is no discussion in the serious adverse event reporting of whether these patients are IGA deficient or not. I know that is one other potential confounding issue.

My question related to the fact that, if you look at your last slide on the adverse event reporting, a lot of it was chest pain and dyspnea.

They also seemed to be even with the prolastin products, a lot of them were more common among the adverse events that are reported.

I was just wondering if anybody had any insight -- this is not specifically maybe to you or to anyone who might have any insight into what those are directly related to.

Are they considered to be allergic reaction? Does it have anything to do with the diffusion of the product into the alveolar space, or is there anything in particular about this product that causes this side effect? I was just wondering.

DR. KHOIE: I am sorry, is your question suggesting that these adverse events are more common in the product prolastin?

DR. DI MICHELE: I was saying, just in general -- Mike made a comment first that it is often difficult to distinguish -- it sort of is a different question than the excipient question, but since there is IGA in all of these products, and there is the potential for some sort of allergic type reactions from the IGA in some patients, that sometimes it is difficult to distinguish product from IGA kinds of events.

Then I was specifically asking about chest pain and dyspnea, as to whether that represented something particular with this product, since we know that this compound basically diffuses into the lung tissue. I just wondered if anybody had any insight as to what those side effects were due to, or what those adverse events were due to. It is a general question.

DR. KHOIE: There is the confounding factor that the indication for the product is clinical emphysema in patients. So, these patients have severe underlying lung disease.

They do also include in their labels the fact that there is the potential for volume overload. That is one potential explanation.

Reading through these reports, unfortunately, i can't summarize all of them for you and say what most of them were likely attributable to.

DR. PIERCE: If I could give a brief follow up to that question, we did look specifically at the chest pain events, to evaluate whether ischemia appeared to be responsible for a proportion of those.

That did not appear to be a worrisome concern, but we did ask the manufacturer to do a specific analysis also in parallel with us, of the chest pain events.

Many, although not necessarily the majority, of the dyspnea cases appeared to be compatible with acute hypersensitivity reaction.

DR. EPSTEIN: Dr. Khoie, thank you for this summary of the FDA data. You very carefully gave us all the limitations and caveats, but didn't actually draw any conclusions.

As I scanned the data, it seems that the relative number and frequency of the report of adverse events in aralast seemed to mirror the relative proportions in prolastin. Is that, or is that not a fair conclusion?

In other words, if you scan the prolastin column and you just ask, what are the predominantly reported findings, you discover that those are the ones where you show early signals with aralast.

Is it the conclusion of the group that there was no apparent difference in the type and profile of reports, given the limitation of the number of reports?

DR. KHOIE: We really felt that we couldn't make a firm conclusion in terms of being able to compare the safety profile of these three products, just based on the surveillance data. It certainly is possible, but I just wouldn't feel comfortable making that conclusive statement.

DR. KUEHNERT: I would just agree with you. I think you mentioned that it wasn't possible to get denominator data.

I know that EARS and Medwatch is sentinel reporting for numerators but, without denominator data, you just can't make much of anything, not just about rates, but also about underlying demographics and patient conditions for people getting these different products.

I would say that, as per Dr. Epstein's question, it would be very difficult to try to make those comparisons without that information.

DR. SZYMANSKI: About the instructions of infusing these products, are you supposed to give them antihistamines, and are these reactions related to the fact after being treated prior to infusion or what?

DR. KHOIE: In terms of the symptoms?

DR. SZYMANSKI: Like people, when they give plasma, they usually give them antihistamines and tylenol and stuff. Do you give that kind of premedication in these products?

DR. KHOIE: I will let my colleague in OBRR answer that question.

DR. PIERCE: I think some of the pulmonologists here might answer that question also, but I can tell you that in the package insert there is no recommendation for premedication for any of the alpha-1 PI products.

DR. DUFFELL: You mentioned before that we don't have a lot of information about other co-morbidities or medications that these patients were on.

It looks like, just given that, it is kind of hard to draw some firm conclusions on this, although the data suggested that there is a correlation here.

As part of the clinical trials, was there any challenge, re-challenge testing done of these adverse events, or did they just occur by happenstance in the clinic, that were ever reported on any of the specific events?

DR. KHOIE: I only have data from the actual surveillance. Regarding the clinical trials, I would also have to defer that to OBRR.

DR. PIERCE: I do not recall any challenge to better ascertain the causality relationship between adverse events reported during clinical trials of these products. If the manufacturers recall any such events, they would be free to also comment.

DR. DI BISCEGLIE: You said you were going to tell us about serious adverse events, but I didn't see any listed or broken out. Were there serious adverse events, and could you give us some idea of what those were?

DR. KHOIE: Yes, there were -- the serious adverse events that I had on the table, I do have a list of them broken down.

Again, I wasn't able to make any nice summarizing comments on them. I could tell you -- I could go through them with you. I would just have to go through the slides. I have the details if you would like that.

DR. DI BISCEGLIE: Not if there is no clear pattern.

DR. ALLEN: It is five minutes after 11:00 by my watch, which is about two minutes fast. We are half an hour behind. At this point, we have been sitting for three hours. We have got three more presentations which adds up to, if everyone is on time, another 30 minutes of presentations. I am going to call a 10-minute break right now, and please be back in 10 minutes, and ready to start talking again.

[Brief recess.]

DR. ALLEN: Dr. Pierce, would you go ahead with your presentation, please?

Agenda Item: Post-Marketing Study Commitments for Licensed Alpha-1 PI Products - Rationale.

DR. PIERCE: I am going to make a brief announcement with relation to the first topic on the agenda earlier today.

Just very briefly, I just wanted to let people know that I was informed that this morning FDA posted a health alert on the CBER web site regarding this product device interaction with maltose and galactose containing products, and xylose interfering with the GDHPQQ glucose testing methods.

I wanted to mention that, on that health alert on the FDA web site, there is a complete list of the GDHPQQ testing method products, as well as a preliminary list of those products we have been able to identify to date that contain the sugars that can interfere with the measurement of glucose by that test system.

So, let me now proceed to the topic at hand. I am going to be talking about the post-marketing clinical studies of alpha-1 PI products.

It will be useful for the committee to keep in mind two of the three questions that they are charged with today. I will again read the first one.

Based on the differences in primary structure of alpha-1 PI, does the committee have comments or recommendations regarding the adequacy of requested and planned post-marketing commitment studies, to evaluate the longer-term safety and efficacy of alpha-1 PI products, as measured by specified clinically meaningful end points. Second, will the committee have comments on the adequacy of the proposed safety monitoring programs.

So, you have heard a little bit about, from Dr. Brantly, the pivotal trial for aralast. In considering the newer alpha-1 PI products, the pivotal trials employed 14 to 30 subjects, randomized to the test article. There was one drop out in the case of the aralast study.

In the cast of zemaira, there were additional non-pivotal trials done, so that the total patient exposure was larger than 30.

These trials were based on the biochemical surrogate end points, antigenic alpha-1 PI or AAT, and neutrophil elastase inhibitory capacity, which is a functional measure.

There needed to be a maintenance of the trough steady state serum levels greater than an arbitrary threshold, and not inferior to that of prolastin.

Levels in bronchopulmonary lavage fluid were expected to be significantly greater than at pre-treatment, and we also had six months or greater safety data, including a measurement of alpha-1 PI antibodies.

So, the sponsors of all three licensed alpha-1 PI products were asked to conduct post-marketing studies to help evaluate efficacy, not just of biochemical end points, but clinically meaningful end points directly related to emphysema.

Now, the first study that was conducted was that with prolastin, and that was an epidemiology study. In contrast, the manufacturers of the newer products have been asked to conduct placebo controlled or parallel controlled, randomized controlled trials.

In the case of zemaira, the manufacturer has agreed to conduct a placebo controlled randomized controlled trial, and there is an approved protocol for that study.

In the case of aralast, the company has agreed to conduct a randomized controlled trial, and the final protocol is pending.

Let me turn back the clock to the completed NIH registry study, the epidemiology study. At the time that prolastin was approved and that this study was conceived, it was considered that a randomized controlled trial was unfeasible.

However, due to its non-randomized nature, the design was not capable of validating or confirming that the PK surrogate end point translated into a clinical benefit.

There was no testing in that study of alpha-1 PI blood levels, and there were large numerical imbalances between the groups of subjects or patients that were receiving augmentation therapy with prolastin, and those that were not. This study was reported in 1998.

The mean findings of the study were that there was a decreased mortality observed among those who received alpha-1 PI compared to those that did not, with a relative risk of .64 and the confidence intervals that you see. That was significant at the .02 level.

For the overall study population, there was no difference in the rate of FVE-1 decline between those on prolastin and those who had never taken prolastin during the trial, except in a post hoc subgroup of individuals with moderate emphysema.

So, several experts have recommended randomized controlled trials to establish the efficacy of alpha-1 PI augmentation, not just biochemical efficacy, but clinical efficacy.

The group that did the NHLBI registry study with prolastin said that definitive conclusion will require a randomized controlled trial.

This was repeated in another publication in the year 2000. A European group of Luisetti said that there was an urgent need to conduct these trials in 2002.

One of the principal authors in the NHLBI registry study wrote also, in 2002, that randomized trials of this were needed to evaluate efficacy, in view of the costly and inconvenient nature of augmentation therapy.

Well, there has been a single randomized controlled, placebo controlled, double blind trial, that was published in 1999 by Dirksen, et al.

This trial employed 56 subjects, and it had the findings -- it used a French alpha-1 PI. The findings were, at the end of the study period, there was a non-significant trend for the alpha-1 PI randomization group to be worse than the placebo group by FEV-1, with a p value of the trend of .2. So, that turned out an opposite of what one would have hoped.

However, they also looked at efficacy in a different way, by CT, which studies have indicated is a more sensitive measure.

There, by one way of looking at the CT outcome, they obtained borderline significance of a p value of .07. By another way of looking at the CT, they had a p value of .2, the same magnitude as the p value of the adverse trend, when looking at FEV-1.

Now, they did a retrospective power analysis and suggested that, with 130 subjects, you could do a CT end point randomized controlled trial and, if the magnitude of the treatment benefit was the same as what they saw in this trial, that would achieve statistical significance, and this looked to be something feasible.

So, the sponsors of all three U.S. licensed studies have either conducted, or agreed to conduct, these post-marketing studies.

What we have asked the manufacturers of the newer products to do is to conduct their investigations in two stages.

The first stage I will get into in a moment, but the end points that we suggested they consider looking at were serial end density changes of ICT, pulmonary exacerbations of COPD -- and there is the suggestion that this therapy may be effective in reducing those, but there are no good hard data in that regard, an important question -- serial pulmonary function testing, and mortality would also be appropriate to examine.

Now, the first stage objective, of the first stage of the investigation, would be to estimate a magnitude of the treatment effect, and to assist in sample size determination for an adequately powered stage two follow up study.

The objective of the second stage would be to actually answer the question, is there substantial evidence of safety, clinical evidence of efficacy -- rather -- and to also gather additional longer-term safety data.

Details with respect to the stage one that we asked companies to consider were to use a randomized controlled parallel masked design, to have a minimum of 60 subjects -- they could, of course, choose more -- that would be 30 per group.

We wanted to be flexible in terms of certain aspects of the design, recognizing the difficulties of employing a placebo control here in the United States, although we believed that a placebo control was feasible in the many countries where alpha PI, to this date, has never been licensed.

So, we indicated that a sponsor could choose a dose controlled design, where they would look at the efficacy of a standard dose in comparison to a higher dose, and there is some theoretical reason to believe that higher doses might be more effective. Also, the placebo control option.

We recommended that the stage one of the investigation be a minimum of one year duration to avoid seasonal bias when looking at the pulmonary exacerbations, which we thought was an important end point.

We also believed it important to measure baseline and steady state antigenic and functional alpha-1 PI blood levels, and we anticipated that such a trial could also have a longer term follow up.

With respect to the conduct of the stage II investigation, we recognized that this could be contingent on the outcome of the stage one.

A strongly positive outcome in stage one might obviate the need for stage II study, also depending on if the size of the stage one study had been boosted from the minimum that we requested.

The performance of the stage II study could be dependent on the amount of products available and earlier, before there were three products on the market, there were intermittent shortages of prolastin, we were told.

It might be dependent on the availability of subjects, the number of subject follow up years needed and, in the event that manufacturers got together and conducted a factorial design study, there might be a contingency on the willingness of other manufacturers to participate.

So, now I will turn to the four post-marketing studies, or phase IV studies, or phase III-IV studies that Baxter has committed to conduct for their Aralast product.

The first one is a stage I efficacy randomized control trial. This will use an investigational version of the alpha-1 PI product, as opposed to the version that has been discussed today.

They have committed to conduct a stage II efficacy, randomized control trial with the same investigational version of the alpha-1 PI intravenous product, notwithstanding those contingencies that I alluded to earlier.

They have agreed to conduct a repeat bronchial alveolar lavage study to answer some residual questions with respect to the data from their study for licensure. This also will involve an investigational version of the alpha-1 PI.

The fourth study is an uncontrolled safety study that is actually an amendment to an ongoing non-FDA regulated post-marketing study, if you will.

This will use a current version of the alpha-1 PI, the one you heard about today, with that C terminal lysine truncation in the majority of the molecules in the vial.

This will be a minimum of 60 subjects. It will last two years, and Baxter will be looking at alpha-1 PI antibody formation, hepatic and renal function, adverse events, and quality of life measures.

Now I am going to talk about the post-marketing studies that ZLB Behring is conducting with their Zemaira product.

They have an approved protocol to conduct a two-year, 100-subject placebo controlled, multi-national, parallel group stage one study, whose primary objective is to look at the effect of Zemaira on the progression of emphysema, assessed by a decline in lung density as measured by CT.

Because this is a stage one investigation, this study is powered to only detect a trend of at least a p value of .2.

The secondary objectives are to assess the effect of the treatment of alpha-1 PI on the number and duration of pulmonary exacerbations, also on FAV-1, diffusion capacity of carbon monoxide, alpha-1 PI levels, exercise capacity, mortality, BMI, safety, quality of life, and anti-alpha-1 PI antibodies.

So, in conclusion, I can say that definitive clinical efficacy data are still lacking for alpha-1 PI products.

Newer data suggests that clinical efficacy studies now are feasible, which was not the thought at the time that the innovator product was licensed in the late 1980s.

Sponsors of all newer intravenous alpha-1 PI products are asked to conduct, or have been asked to conduct, two-stage post-marketing investigations of one or more clinically meaningful end points.

Longer term, two-year safety data are being collected for both Aralast and Zemaira, the two newer products, in 50 to 60 subjects, to include alpha-1 PI antibodies.

Again, I hope that my remarks have been useful to the committee in wrestling with these two questions that we posted earlier, and any advice that the committee has with respect to recommendations we should be making, in addition to the companies, regarding the design aspects of their trial or any safety parameters that should be looked at in addition to those that I have outlined would certainly be appreciated. Thank you.

DR. ALLEN: Thank you, Dr. Pierce. We have got two more FDA presentations. I would really like to limit the questioning now to just clarification or direct referral to the slides, where you want them back up, because we will have a chance for discussion later.

DR. KUEHNERT: Just a clarification. You mentioned limited studies, but did you mention anywhere -- maybe I missed it -- ongoing collection of safety data that would be asked of the companies? That has no time limit, for ongoing data.

DR. PIERCE: That is a difficult question to answer, in some sense. We have an approved protocol in the case of the zemaira product. To my understanding, that trial has not yet initiated, but may shortly.

We have the agreement in principle to conduct the clinically meaningful end point study with additional safety data from the manufacturer of aralast.

In the case of both products, there is ongoing active surveillance, not just passive surveillance, with respect to adverse events. I believe that more information may be available in this regard.

In terms of the formalized post-marketing commitment to obtain information on aralast, on adverse events over a two-year period and 60 subjects, we are awaiting a finalization of the protocol for that, but that is really an amendment to an ongoing study that was voluntarily initiated by the firm, and their contractors. I hope that answers your question.

DR. SCHREIBER: Does prolastin not have to do a clinical trial?

DR. PIERCE: Prolastin, at the time that it was licensed, there was an agreement to conduct a phase IV study that would look at lung function.

The thinking at that time was that a randomized control trial was not feasible in terms of the numbers, that you just wouldn't be able to recruit enough patients.

So, the sponsor substituted the NHLBI registry study for the trial that they had agreed to conduct, and FDA accepted the substitution of that. We have not requested that now Talecris conduct a randomized trial with prolastin to date.

DR. SCHREIBER: Why is that?

DR. PIERCE: It has been something that has been discussed.

DR. SCHREIBER: The other two products are not licensed?

DR. PIERCE: We have three licensed products. The prolastin was licensed in the late 1980s, and then the other two products, including the one that we have been emphasizing today about the isoelectric forms --

DR. SCHREIBER: It was just surprising to me that three licensed products, you can require two of them to go back and do clinical trials and the other one gets off scot free.

DR. PIERCE: Talecris will actually be making a presentation during the open public session today, and they may wish to comment further on information of a controlled nature with respect to their product, and the potential for evaluating some of these more immediately clinically pertinent end points.

DR. ALLEN: Okay, other questions or comments? All right, the next FDA presentation has two speakers. it is under one title topic, however, licensed biological products with structural heterogeneity, with known structural modifications. Dr. Andrew Chang will be the first presenter.

Agenda Item: Licensed Therapeutic Protein Products with Known Structural Modifications.

DR. CHANG: Thank you, Mr. Chairman, good morning. Before I start my presentation, I just want to repeat a statement that Dr. Andrew Shrake made in his introductory speech earlier this morning.

That is, the biological products in the United States, at least, are licensed on a case by case basis, and based upon the standard specified in our regulations.

My name is Andrew Chang. I am associate director in the division of hematology, office of blood, CBER. As Mr. Chairman pointed out, this presentation will be divided into two speakers.

Because of the limit of time today, I selected two case studies that related to plasma derived products, and Dr. Kurt Brorson from the division of monoclonal antibodies, office of pharmaceutical science, CDER, will present some of the case studies that are related to recombinant and monoclonal antibodies.

He also will present some of the strategies that can be used to maintain the quality of the biological product and some strategies that can be used to assess the impact of the heterogeneity.

There are at least two major factors that can contribute to the heterogeneity of the biological products. The first one is biosynthetic processes used by the living organism.

This factor mainly actually applies to recombinant products as well as monoclonal antibodies, and has a minimal, but not zero, impact to plasma derived products, for the fact that plasma derived products, normally are manufactured from a large plasma pool. So, any effect from living organisms will dilute it, because of a large plasma pool.

The second major factor contributing to the product heterogeneity is involved with the manufacturing process, and also the storage conditions that are used for a drug substance and a drug product.

Now, how to control structural heterogeneity for a biological product? There are two major elements that should be considered to control heterogeneity.

First, the sponsor or applicant has to demonstrate the consistency of the heterogeneity pattern of commercial lots with that of the lots used in preclinical and clinical studies.

Second, to assure lot to lot consistency, the content and extent of this heterogeneity should be characterized and controlled.

Here is a list of the potential protein modifications that can contribute to the heterogeneity of biological products.

This is not meant to be inclusive, but rather, examples of the protein modifications that can happen to a biological product.

These include the primary sequence heterogeneity, the post-translational modifications, and the high order of the protein structure.

The first case study I want to present to you today is a charge analysis of unlinked oligosaccharides from factor VIII standards, recombinant and plasma derived factor VIII products.

Most of you, I am sure probably all of you know, that factor VIII, an activated form of factor VIII, is a very important cofactor for intrinsic coagulation pathways.

Patients with defects in the factor VIII molecule, or absence of the factor VIII, will have a bleeding disorder, and we call them hemophilia A patients.

Factor VIII is a multi-domain protein that has estimated 24 unlinked polyglycans, oligosaccharides. It is worth it to point out that the majority of the polyglycan is located in a domain called the P domain.

Factor VIII undergoes activation in vivo and becomes the activated form of factor VIII. We have, in the United States, we have multiple licensed factor VIII products, including recombinant and plasma derived factor VIII products.

One of the licensed products, so-called P domain deleted factor VIII, the difference between the P domain deleted and the four lanes(?) factor VIII is the absence of the P domain.

Dr. Schilow et al from the Port Eric(?) Institute, reported their study which I presented to you today.

What they did is to look at a compilation of oligosaccharides among different licensed factor VIII products, including recombinant and plasma derived factor VIII, as well as some national and international standards.

What they did is, they isolated the oligosaccharides, and then separated them according to, in this case, the sialic acid charge.

So, they separated them into mutual glycans, monocytic glycans, diactyic glycans, triacytic glycans, and a tetracytic glycans.

Then they did analysis on the different products, and what they found is that, as they reported, there are at least three groups.

You can divide the licensed or some of the standards into three groups according to the differences in composition of oligosaccharides.

The first group, the group is a group of the factor eight products that has a high level of the diacytic, triacytic and tetracytic oligosaccharides, greater than 70 percent of the oligosaccharides containing the di, tri and tetrascytic glycans.

The second group is a group of products that have greater than 70 percent of the oligosaccharides containing mono and diacytic glycans.

If you look at the product that is approved, you can find that not only this grouping applies to -- this grouping applies to not only recombinant, but also the plasma derived factor VIII.

For instance, in this case, the plasma derived factor VIII, which is categorized as group two, which has greater than 70 percent of the mono and diacytic glycans. The last group they identified is a B domain delayed product, named Refacto, and it has the unique oligosaccharide composition.

In conclusion, the charge analysis demonstrated heterogeneity of the glycoforms of the tested factor VIII products.

The second case I want to present to you today is a characterization of five von Willebrand Factor concentrates produced by five different manufacturing processes.

This is actually part of the international study that we participated in for establishing the first international standards for von Willebrand concentrates.

Von Willebrand factor is a complex molecule, and it undergoes serious post-translational modification intracellularly, and it forms multi-molecule complexes. We call them von Willebrand factor multimers.

This protein is rather interesting in a non-sheer stress situation. This multimer has a sphere shape. If you provide a sheer stress to this model, they will undergo shape change. In this case it is really a fiber type of the multimer, and it depends on how much sheer stress you provide. They have different conformations.

We looked at five candidates which are either factor VIII concentrate or von Willebrand factor concentrate, as the five candidates for selection of the first international standard for von Willebrand factor concentrates.

We looked at a multimer form among these five different concentrates. As you can see, when you separate different sizes of multimers by agarose gels, and they are detected by western blotting, you see a different size of the multimers on the gel.

The first line is the international plasma standard, and the second line is a plasma by itself. Then the rest of the five candidates.

If you look at a one percent gel, they have, for instance, the C2 log of the high molecular weight multimers. If you are running the gel by a high concentration of the agarose, you start to see some differences at low molecular weight multimer forms.

If you look at the function and compare the specific activity among different concentrates, you can actually also separate them into three groups.

For instance, the C-1 candidate has specific activity -- risto setting(?) co-factor versus antigen level close to one, as well as the collagen binding activity versus antigen level.

C-2 actually behaves quick differently from C1, and the specific activity is close to, or below, 50 percent. The rest of the three candidates has specific activity close to 70 percent.

In conclusion, the characteristics -- in this case, a multimer form -- and the specific activity of von Willebrand factor concentrates, depends on the manufacturing process.

Now I am going to give the podium to Dr. Kurt Brorson for his presentation.

Agenda Item: Licensed Therapeutic Products.

DR. BRORSON: Good morning, and thank you for your attention. As Andy mentioned, I am in CBER's sister center, the Center for Drug Evaluation and Research.

In the Center for Drug Evaluation and Research, we regulate, of course, the small molecule drugs that everyone is familiar with but also, in addition to that, the majority of the biotechnology products.

All of the CMC reviews are conducted within two offices of CDER, the Office of Biotechnology Products, and the Office of New Drug Chemistry.

Within the Office of Biotechnology Products, we have two divisions, the division of monoclonal antibodies, which obviously reviews monoclonal antibody applications, as well as the division for therapeutic proteins, which review many other types of biotechnology products.

I am in the division of monoclonal antibodies, and I am here to give you my perspective on the types of heterogeneity that we see, both in monoclonal antibodies and in therapeutic proteins made by recombinant DNA technology.

The first thing to remember about biophramaceuticals is that they are quite complex, and they have much more potential heterogeneity than small molecule drugs.

To really illustrate that, here is a structural model of basically one third of a monoclonal antibody, just the FAB region, compared to a statin, a typical small molecule drug.

You can see that even one third of an antibody is much larger than a small statin molecule. The source of this heterogeneity comes from -- there are a variety of different types of heterogeneity, and they can be broken down into two types, cell culture related heterogeneity and stability related heterogeneity. So, let me first describe cell culture related heterogeneity.

The first thing to keep in mind is that fermentation of cell culture in a bioreactor, for example, is really an artificial process.

You have cells growing in a bioreactor, in a media that is usually a complex mixture of chemicals and nutrients.

Relative to plasma and other natural sources, that environment is much more artificial. Because of that, some cell culture related heterogeneity to protein product can result.

In fact, the bioreactor conditions themselves, the pH of the media, the temperature of the media, how long you run the bioreactor, whether you run it in batch mode or continuous mode, can impact a variety of things, including the glycosylation, certain charge variance of the product, as well as some types of substitutions such as a leucine or norleucine substitution in the final product.

In addition, there is clonal variation between cell lines, of course. The cell line that is used to produce the product actually can impact things like adduct placement, the number of adducts on cysteine molecules, for example, folding or mis-folding of the product, as well as cysteine pairing. So, all of these cell culture related issues can lead to heterogeneity in a biotechnology product.

There are also stability related issues with heterogeneity. Again, it is important to remember that a pure, high concentration protein solution is an artificial system compared to, for example, proteins in cells in bodily fluids. Therefore, you are stressing the proteins to some extent, just by having a high concentration formulation.

However, I should note that many of the plasma products face the same issues as recombinant products. So we actually are not that different from the products that you regulate here, in terms of stability related issues.

Typical stability related heterogeneity can include things like clipping, aggregation, deamidation where asparagine can turn into aspartic acid, or glutamine can turn into glutamic acid, or the loss of epsilon amino acids on lysine. That will cause a lot of charge variation.

Finally, another very common stability related form of heterogeneity is in oxidation, where methianines are turned into methiadine sulfoxide.

Now, for monoclonal antibodies, there are two major sources of heterogeneity. The first source is right here, in the glycosylation or the carbohydrates that are attached to the FC region.

Essentially the major form of heterogeneity here is the number of galactose molecules or moieties that are attached to this GLIC neck(?). So, you can have either zero, one or two galactoses here.

The second form of heterogeneity that is very common in monoclonal antibodies will exist right here at the C terminus of the FC region.

Now, keep in mind that antibodies are dimers. So, there are actually two C termini here. That heterogeneity basically is whether or not there is a C terminal lysine at the end of either zero, one or two of the proteins that form the antibody dimer.

This has been our experience with 20 or 25-plus monoclonal antibodies that we have licensed so far, plus a variety of other antibodies that are under IND.

So, we have come to expect this type of heterogeneity, and we have developed strategies to deal with it.

This right here demonstrates the C terminal lysine heterogeneity. This is not an IEF gel. This is actually a substitute method called the weak cation exchange HPLC column.

It basically measures the same thing. It measures charge variance, and this is a chromatogram of such a column.

You can see here that you can distinguish between the antibodies with zero C terminal lysines from the antibodies with one terminal lysine, from the antibodies with two of them.

We see this repeatedly with monoclonal antibody products. The way this is approached is that manufacturers will set acceptable range either for each species or for the three of them combined.

They can be measured by various techniques, IEF gels, which you have seen extensively this morning are merely one technique. This is the weak cation exchange column, which is also an acceptable method to look for charge heterogeneity.

I would imagine that there are many other methods, for example, capillary isoelectric focusing, that could do the same thing.

It has been our experience that this pattern that we see repeatedly doesn't seem to impact the potency or the safety profile of the product.

Now, of course, we do see unacceptable heterogeneity. Normally the manufacturers will run a stress study, where you store the antibodies at perhaps 40 degrees C, 45 degrees C, in order to look for degradation pathways.

You can see, here are two examples of stress induced heterogeneity. In this experiment here, the antibody was clipped into an FAB fragment and FC fragment which, again, could be resolved on this weak cation exchange column.

In this other stability study, the antibody underwent deamidation. So, basically the charged variance here became even more acidity and bound to the column more tightly, and you can see a pattern like this.

These would clearly be unacceptable for a licensed product, and manufacturers will set stability specifications looking out for this type of pattern.

If they are out of specification -- and clearly these are out of specification -- it actually can impact potency of the product.

I believe that this antibody was tested here, and it had about 50 percent of the potency of the intact antibody.

The safety is a good question. We would never allow an antibody that looked like this to be infused into a patient, even in a clinical trial. It just is deamidated too much, and it may actually end up being immunogenic.

This actually probably would lose some potency, but probably wouldn't pose a safety issue, merely because these are really not neoantigens, or neomolecules for the patients. They are just merely half of the active ingredient in each of the species. Again, we would not want this to be injected into a patient.

I discussed heterogeneity with our sister division, the division of therapeutic proteins, and they were able to think of three cases where they saw significant amino acid variation.

The three cases that they recalled -- when I talked to the deputy division director there, so this probably constitutes most of what they have seen in terms of heterogeneity -- only one of the cases actually had an impact on potency.

They saw one case of protein terminus heterogeneity traced to a metalloprotease, and one case of product clipping. In both cases it had minimal impact on the potency of the product.

They did have one case where N terminal glutamine cyclyzation occurred. Interestingly, the cyclyzed form had a two-and-a-half-fold greater potency.

Again, this is something that can get picked up in the potency assay itself, which is a normal part of lot release for all products.

So, what does the office of biotechnology products do to keep a handle on heterogeneity? First, obviously, there is testing at lot release and stability.

The strategy for lot release testing is to employ a range of assays to look at as many product quality attributes as possible at the given time.

Obviously, we will look for charge variance using IEF or weak cation exchange chromatography. The sponsors will also look for amino acid sequence variation, oxidation, using internal mapping or C terminal or N terminal mapping, that can be used to look for clipping or aggregate formation.

Mass spectrometry is highly sensitive to molecular weight changes. It is not always used in lot release because the equipment is expensive, but it can be used for that purpose.

Then there is a whole variety of specialized assays that will look for either carbohydrate changes or, in the case of isoquant, it looks for deamidation.

Acceptable ranges for the heterogeneity is based set on basically a combination of clinical experience and manufacturing experience.

It is a combination of what is acceptable in the clinic, as demonstrated by the clinical trials, compared to what is a manufacturing process capable of.

The second strategy is formulation. You can formulate your proteins to minimize heterogeneity, and if heterogeneity cannot be avoided, the goal is to control it and answer three questions.

First, does the heterogeneity impact the active ingredient potency. Does it impact bioavailability and, finally, does it impact immunogenicity.

This can be addressed by a variety of approaches, including careful examination of data from the potency assay, TK studies and, in extreme cases, sometimes immunogenicity studies might be warranted. Thank you.

DR. ALLEN: Thank you. At this point, unfortunately, we are almost an hour behind. I want to make sure that we have got adequate time for committee discussion. Are there burning questions for either Dr.Chang or Dr. Brorson from their presentations?

If not, we will move to the open public hearing. Again, I would like to remind the speakers that have asked to speak to please appreciate the time constraints on the committee. Give us your presentation very succinctly, make your important points, and we will then move on.

Agenda Item: Open Public Hearing.

DR. ALLEN: Let me read the open public hearing announcement for considered matters meeting. Both the Food and Drug Administration and the public believe in a transparent process for information gathering and decision making.

To ensure such transparency at the open public hearing session of the advisory committee meeting, FDA believes that it is important to understand the context of an individual's presentation.

For this reason, FDA encourages you, the open public hearing speaker, at the beginning of your written or oral statement, to advise the committee of any financial relationships that you may have with the sponsors and, 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.

If you do not choose to address this issue of financial relationships at the beginning of your statement, it will not preclude you from speaking.

The first speaker who has asked to speak is Miriam O'Day with the Alpha-1 Association, Annapolis, Maryland.

Agenda Item: Remarks by Miriam O'Day.

MS. O'DAY: Thank you very much I am Miriam O'Day and I am a contract consultant for the Alpha-1 Association and also for the Alpha-1 Foundation, and my statement today is on behalf of the Alpha-1 Association, which represents the community of individuals with alpha-1 atrypsin deficiency, the majority of whom are frequent and life-long plasma therapeutic product recipients.

The association remains concerned about the differences in the three augmentation therapies that are currently available for the lung disease associated with alpha-1.

The association contacted the FDA via telephone in October of 2004, and Dr. Shrake mentioned that on his slide. He had that noted as the Alpha-1 Foundation. It was actually the association.

Then we also contacted -- the association again - in January 2005, in writing, when these issues were brought to our attention once again by concerned physicians.

We continue to hear some concerns expressed, and mainly that there remain a lot of unanswered questions, and we stated some of those questions in our correspondence in January.

The primary ones are, is the variance an indication that the product may not be functioning as anticipated, or otherwise causing harm.

Have physicians been contacted, so that they are aware that they should be watching for adverse events that should be reported to the FDA, and has comprehensive patient follow up been conducted on individuals receiving Baxter's aralast, to determine the clinical significance of these variances.

We remain concerned about the lack of information being given to patients and health care providers about these products, that may leave these variances and adverse events unreported and undiagnosed.

We have entered an era when the blood supply is significantly safer than it has been in the past. In fact, manufacturers frequently use safety as a marketing tool when talking or interacting with consumers.

As individuals become much more educated about blood safety and certain methods of viral inactivation, they have also become more savvy about the impact of infusing blood products on a regular basis.

The FDA should be working with the manufacturer to establish clear protocols for post-market surveillance that will be shared with the consumer community.

The current passive system of adverse event reporting is not a viable evaluation method of product safety, as we heard this morning.

We remain keenly interested in the details of additional ways that the FDA will measure the clinical implications of the variances in aralast.

We thank you for the opportunity to address the committee, and we would like to leave you with the following three recommendations.

One, a dear doctor letter should be issued by the FDA to clinicians treating individuals with alpha-1 a-trypsin deficiency.

Baxter should be required to conduct comprehensive post-market surveillance to determine the long-term clinical effects of said augmentation irregularities, and make such information available to the public.

Number three, that the comprehensive post-market data should be collected and analyzed prior to granting authority for Baxter to increase production of aralast. Thank you very much for your time today.

DR. ALLEN: Thank you, Ms. O'Day. Next, Ms. Sally Everett also with the Alpha-1 Association.

Agenda Item: Remarks by Sarah Everett.

MS. EVERETT: Actually, I am with the Alpha-1 Foundation, and I have no financial relationships to report. My name is Sarah Everett. I am an alpha. That is, I am a person with alpha-1 antitrypsin deficiency.

I was diagnosed in February 1993, and I am here representing the Alpha-1 Foundation. I became a board member of the Alpha-1 Foundation in January 1997.

The Alpha-1 Foundation is dedicated to improving the health, quality of life, and finding a cure for those with alpha-1 deficiency.

To that end, the Alpha-1 Foundation participated both in the recruitment of patients for alpha-1 replacement therapy trials, and the fast tracking of licensure.

For that reason, we have a heightened responsibility to assure both the efficacy and the long-term safety of these drugs.

In February 1993, I joined the seven-year long NIH longitudinal study of prolastin, the first alpha-1 antitrypsin replacement therapy. To this day, I continue weekly prolastin with no side effects.

When diagnosed, I had moderate emphysema. Today, almost 13 years later, I have severe emphysema. I strongly believe, however, that prolastin has slowed my decline, as evidenced by an objective reading of my FEV-1 over time, as well as my ability to travel here to speak to you today.

For those of us with alpha-1 antitrypsin deficiency, by the time of diagnosis, our lungs are already compromised and, in some cases, severely compromised. Alpha-1 antitrypsin replacement therapy is our only lifeline. It is also a lifetime commitment.

On the other hand, since alpha-1 antitrypsin replacement therapy can only be given intravenously, it exposes us to great risk.

When taken together with the already compromised state of our lungs, if something goes wrong, whether lack of efficacy or an adverse event, it can mean life or death.

For that reason, the efficacy and especially the safety of all new alpha-1 antitrypsin replacement therapies is critical. Since this is a life time therapy, detection of potential adverse events in a long-term study is essential.

To that end, the scientific leadership of the Alpha-1 Foundation has strongly recommended long-term post-marketing safety study of all new alpha-1 antitrypsin replacement therapies.

I believe that a copy of a suggested protocol, entitled, continuous safety surveillance study for all alpha-1 patients, has already been provided to you.

On behalf of the Alpha-1 Foundation, I thank you. On behalf of myself and other alphas, our lives are in your hands. Thank you.

DR. ALLEN: Thank you, Ms. Everett, for a very important statement from the foundation and on behalf of yourself personally.

Our next speakers, we have two presentations from industry. I have got Talecris Biotherpeutics, either Dr. Robin Huff or Dr. Barbara Merrill.

Agenda Item: Remarks by Barbara Merrill.

DR. MERRILL: I have a single slide. I am Dr. Barbara Merrill from Talecris, as you stated, and we have seen a lot of IEF profiles this morning.

Obviously, we didn't see all the presentations when we submitted our -- basically, what we are showing you here is a bit of a twist on what previous speakers have presented.

This is actually a western blot. So, it is isoelectric focusing which you have already heard a great deal about.

After we did the isoelectric focusing, the proteins are transferred to a membrane, and then probed with antibodies that are specific to alpha-1 PI.

So, all that you see is a signal to alpha-1 PI. We do this in part because we know there are other proteins in prolastin, and we wanted to just do a specific comparison of alpha-1 PI content between our product, prolastin, and a plasma pool.

So, the second point I want to make about this slide is, unlike Dr. Brantly, who was showing comparison t an M1 standard, which is a single allele, this is comparison to a plasma pool. So, it represents the lower level variance that you might see in a normal human population, which is also what prolastin is derived from.

The key point to note here is that the profiles of prolastin and normal human plasma are extremely similar. I just want to make that point. My colleague will address the clinical question.

Agenda Item: Remarks by Robin Huff.

MS. HUFF: I am Robin Huff, also from Talecris Biotherapeutics, and I wanted to address the question that came up earlier as to whether a randomized controlled trial would be requested for prolastin.

I did want to inform the committee, as well as FDA, that Talecris has voluntarily initiated such a study. It is ongoing. It goes by the name of Exactly, and in fact, it is looking at many of the end points that were previously described by FDA as suggested end points. So, all three manufacturers, it would appear, are on the road to, in fact, exploring these end points.

In addition, I would like to make the note that, unlike the two more recently approved products, because we were approved in 1987, of course we enjoy a greater than 17 year history of safety data that has been collected, that supports the safety of the product, which goes to address the point that, therein lies part of the difference for why, perhaps, the study was not actually mandated, but nevertheless, we are voluntarily initiating such a study.

DR. ALLEN: Can you tell us how many patients you intend to enroll in the trial?

MS. HUFF: I think we are looking somewhere on the order of 70-ish, if I can pus an ish behind it, because I don't want to commit to the specific number. I don't have it handy with me. It is on that magnitude.

DR. ALLEN: Thank you very much.

MS. HUFF: That would be a number to be enrolled. That wouldn't be an established, already enrolled number.

DR. ALLEN: Our final listed speaker is from ZLB Behring, Dr. Otto-Erich Girgsdies.

Agenda Item: Remarks by Otto-Erich Girgsdies.

DR. GIRGSDIES: My name is Otto Girgsdies. I am from ZLB Behring from the clinical research department, located in Marburg, Germany.

I first would like to thank CBER for inviting us to give this presentation. I can make it very short, because most of the points were already mentioned by Dr. Pierce.

We got our approval for Zemaira in July 2003, on the basis of the efficacy and safety data assessed in four clinical studies, and we have included in these studies 90 percents.

This clinical study demonstrated that Zemaira augmented maintains the serum level of API above the threshold of 11 micromolar, and it was also shown that API reaches the epithelial lining fluid, where it has to do its work, and zemaira has been shown to have a similar PK profile as prolastin, which at that time was the only other product on the market.

We got the approval with a post-licensure commitment, and we agreed to perform a randomized, placebo controlled, double blind multicenter study, and we compared 60 milligrams per kilogram body weight of zemaira weekly, IV administration with placebo, and we will include 100 patients in that study.

The observation period for this study is thought to be at least two years, and we also committed to try to extend this study.

The main inclusion criteria is the age between 18 and 60, but probably most of the patients are starting with 35 or 40.

They have to have a for sure diagnosis of alpha-1 deficiency, and they have to have clinical signs of emphysema and FEV-1 between 35 and 70 percent of predictive.

Our primary end point is lung density, and it will be measured by computer tomography, and we asked to perform CT scans at five time points before start of the treatment, and then after three months of the 12 months, 21 and 24 months.

We have chosen end points. Within the last three to five years, CT scanning has improved very much in these patients, and we are thinking tha tit is a most robust end point to show efficacy in such a clinical trial.

Secondary end points, exacerbations, lung function, exercise capacity -- we are thinking of including a shuttle walk test -- body mass index, tolerability, mortality were included, and quality of life. We will use the St. George questionnaire for that.

Pulmonary emphysema is characterized by destruction of the lung tissue, increased area of trapped area, and decreased blood circulation, and this results in a decrease of x-ray attenuation, and I will give you an example of a shift of histogram in a patient and in a normal subject.

Here you can see this is a histogram of a normal subject, and this is the histogram of a patient with emphysema.

Here, on the left side, here is -1000 houndsfield units, which means air, and here is water. In the patients with emphysema, you will get a shift toward the minus thousand parts, and the 15 percentile points we have chosen as effect variable for the whole lung, because it has shown in previous studies the lowest variation.

We have an optimized protocol for that, and it is optimized for density resolutions. We don't have nice pictures, and we have thick slices of five millimeters, and we use low energy for that, and this brings the radiation exposure to the down to 1.0 to .7 mSv effective dose. A normal background radiation per year is about 2 mSv.

We have a validated protocol available for these CT scans, and the same scanner must be used throughout the study. Hopefully it is possible, that it does not break down, and all scans must be acquired according to the final protocol, and this was published very recently by the group of Stoel and colleagues.

We also will perform duplicate CT scans, because it has been shown that repeated scans make it possible to individually correct the volume changes, because this is a major point for making mistakes, and we will have two CT scans, one at full inspiration level and one at functional residual capacity, and this has been shown to reduce the sample size requirement by a factor of about four, irrespective of effect size and power, and the publication is prepared by right now by the group of Stoel, and an abstract has been shown very recently, in September, in Copenhagen, that this works.

Of course, each site will be provided with a density resolution phantom to check for scanner stability with such a phantom, and all CT scans will be analyzed using a validated software package, and a CT Core Lab standardizes and controls the quality of our images, and also evaluates the scans from all the centers. I will hand over now to my colleague, Val Romberg, saying a few words about the residues.

Agenda Item: Remarks by Val Romberg.

MR. ROMBERG: Basically, we have reproduced what Dr. Schwarz talked about for aralast, for zemaira, prolastin and aralast, and found his same results. So, there are really no differences there

I did want to mention, though, that it is important to recognize that, while the cysteine group in alpha-1 exists -- here is kind of a stylized portion -- in vivo, it exists in two forms.

One is the free form or the reduced form, in this case, and the other is the oxidized form where you have a disulfide bond with cysteine.

So, in vivo you have these two forms. Zemaira does consist of strictly this form. However, when it is infused, the expectation is that it then reacts with circulating cysteine and glutathione, and it goes back into equilibrium as per the natural product.

The other thing I wanted to mention is just from a manufacturer's perspective. I know you are considering what to do about potential differences in the structural characteristics of API from different manufacturers.

I think it is important to remember that this is a rare disease. It seems pretty easy to say, let's do another clinical trial, but I can tell you our company has closed two plants in the last two years.

We laid off half the staff at our Zemaira manufacturing site recently, and this is not big pharma. This is the plasma industry, and I would say that financially, and I would ask you to consider that as you are thinking about what your action should be.

DR. ALLEN: Thank you very much. Is there anybody else who wishes to speak during this open public hearing segment?

All right, the open public hearing segment is closed. We now move into the final phase, which is committee discussion and consideration of the questions.

Agenda Item: Committee Discussion.

DR. ALLEN: Why don't we go ahead. Dr. Shrake, will you be putting the questions up for us again, so we have got them on the screen, and then we can have general discussion and address the questions, or it is really parts A, B and C of one question.

The question is: Based on the differences in primary structure of alpha-1 PI, and the concentrations of polymers in the alpha-1 PI products, does the committee have any comments or recommendations concerning, first, the adequacy of the requested and planned post-marketing commitment studies to evaluate the longer-term safety and efficacy of the alpha-1 PI products as measured by specified clinically meaningful end points, second, the adequacy of the proposed safety monitoring programs and, third, any other suggested actions, for example, additional communications through labeling or other venues. These are open for discussion.

DR. NELSON: On that list of adverse reactions, I had real trouble figuring out what was due to the drug and what was due to the underlying condition.

Sometimes, when adverse events are reported, they are reported in three categories, definitely related, probably, or possibly.

I don't know if that is the way these events are reported. Things like dyspnea and hepatitis, that is part of the syndrome.

One way with dyspnea, I guess, is the temporal relationship to the infusion. I had real difficulty interpreting that list, as to what it really meant.

DR. ALLEN: The passive surveillance or adverse event reporting system is not very helpful in this kind of a situation, and clearly it needs to be looked at.

DR. NELSON: I wonder if it was reported in categories, definitely, probably, possibly?

FDA STAFF: Generally, the categories you refer to are in Medwatch reports that come from post-marketing studies, where the investigator is required, under the regs, what they think the assessment of causality is.

Usually, in true spontaneous reports, which the vast, vast majority of these were, it is not explicitly stated.

I can tell you, from reading all of these reports, that in most of them it is not possible to figure out if it is confounding by indication, or if it is some reaction to the product.

In a number of them, the reporter did actually state that they think this is an exacerbation of the underlying disease, and not separately related to the product.

In the case of the reports that were more suggestive of hypersensitivity -- let's say, swollen eyes, shortness of breath, things like that -- it was a little easier to figure out, but you are right, it is a serious limitation.

DR. KULKARNI: I was wondering if there was any data on product switching, when people switch from one type to another one. Is there an increase in adverse event or any impact on its reactivity.

DR. ALLEN: Does anyone have information on product switching?

DR. KULKARNI: I am sure this probably happens all the time.

DR. PIERCE: The design of some of the trials of the newer products include an initial randomization period in parallel, and then the group that was randomized to prolastin would then cross over to gain additional experience with the newer product. Again, nothing stuck out like a sore thumb in terms of that switch point necessarily.

DR. SNIDER: The one suggestion I have is that the surveillance for large numbers of patients for long periods of time -- clinical surveillance, I mean, which is highly standardized in the questions that are asked, seems not to have been listed on the presentation that Dr. Pierce gave us. The chemical and other analyses are fine, but I think we need careful clinical surveillance.

DR. ALLEN: Thank you, and I think that really needs to be done through these prospective studies that the FDA would like to see in place now.

DR. KUEHNERT: Could I just add a comment to that? I brought that up before because it looked like those were time limited. I think what seems to be coming up is continuous surveillance.

DR. CRYER: I think that, in answering question A, that the three companies are doing three different studies, all of which are going to add significant information.

However, it seems to me that the one thing that could be improved upon would be some cooperation among them to design the studies so that they can be compared, that the patients within their own products could be compared.

There is a small number of patients out there. They are all being treated by three different products. Some way to compare the outcomes of those three products would be beneficial, it seems to me. I recognize the economic and political problems of accomplishing that, however.

DR. ALLEN: I think the FDA can take that statement under advisement.

DR. DUFFELL: I was just thinking along the lines of one of the earlier panel members that, given the noise in the data that we do have, and the complexities of these treatments, that I really wonder if we have jumped the gun in requiring a prospective type trial, and whether or not surveillance alone, in some sort of an organized fashion, would accomplish the type of data that we need to determine whether or not a true randomized controlled study is needed.

That is just my comment. It looks to me like it might be premature to even require the study. Surveillance alone, in an organized fashion, might accomplish what we needed.

DR. ALLEN: Could you explain a little bit what your concept of surveillance in this is, whether it involves tracking of selected patients in a trial sort of environment, or what exactly would you include in it?

DR. DUFFELL: Kind of a quasi-trial, but really more of a post-market adverse event reporting type schema that was thought out in a prospective way, that we elicited the information that we think may be underlying here.

I mentioned before the idea of a challenge re-challenge type opportunity. I think that would also, under controlled circumstances, especially in a clinical setting, could be done safely to determine whether or not there is a real causal relationship here.

DR. NELSON: That wouldn't give you efficacy data.

DR. DUFFELL: I am presuming efficacy has to have been established in order for these products to have originally been cleared to the market.

DR. NELSON: One of the presentations suggested that it hasn't or, at least depending on the end point, that it is still questionable.

DR. KLEIN: In reading the first question, after listening to all these elegant biochemical presentations this morning, I didn't see any evidence that these minimal changes in primary structure or concentration of polymers was related to anything clinically at all.

I guess the good news to me is that, over the course of some 18 years and several manufacturers, the safety data of these infusions, given that they are infusions of large concentrations of proteins, seems to be pretty good.

The bad news, I think, is that, despite the fact that we have seen 18 years, I don't see a shred of evidence that there is any efficacy here at all, unless I have been missing something.

I think really, at this point in time, we need some efficacy data. The phase IV studies, I think, need to be designed in order to get us more than the surrogate markers of circulating protein or the presence of protein in BAL. It really has to do something with the function of the lung and the function of the human being.

DR. SZYMANSKI: I agree with Dr. Klein's opinions, and I think we really should have efficacy data. However, I think this might be difficult, because if this protein is needed from very early on to protect the lungs, and if you do not give therapy until you have reached emphysema of a certain degree, how can it get cured?

You can't really make new lung tissue. So, therefore, I think that the studies might be disappointing. An ideal study would be to start treating very early on before emphysema has developed.

DR. BISCEGLIE: Actually, Dr. Klein said very elegantly what I wanted to say. I really have not seen anything to be too concerned about in this micro heterogeneity in terms of an impact on patient safety.

I would like to see a little more data on immunogenicity. We heard just a statement that it hadn't altered it, but we didn't see any data.

I think the big issue in my mind is, there is no evidence of efficacy. We have heard a very passionate plea from patient advocates to reassure them of safety, and I think we should do that.

The horse is kind of out of the barn, these things are licensed. What are we going to do, if we do these studies and can't prove any efficacy.

I think we have to do it. I think we have to try to show that these things work and then, if they don't work, then we are going to be stuck.

DR. ALLEN: I think the point was made earlier in the presentations about the need for an adequate population sample in order to do the efficacy studies.

DR. LAAL: It is not clear to me if the different lots are being standardized on the basis of protein concentration or on the basis of enzyme activity, catalytic activity.

That could be one of the reasons why we are not seeing a lot of efficacy, because we are monitoring for the wrong thing. We are not monitoring for the function. I didn't see that data at all today.

DR. DI MICHELE: I just wanted to say that I agree with a lot of what has been said with respect to the potential relative importance of safety and efficacy data that has to be gathered.

I also would like to say that the issue that was brought up by one of the industry representatives was the fact that this was a rare disorder, and that approaches to rare disorders really have to be considered separately.

I know that we are discussing this a lot in terms of the rare plasma protein disorders for coagulation, the rare coagulation deficiencies.

In many cases, the issue of surrogate markers to get proteins licensed followed by post-marketing surveillance to really look at other markers of efficacy and safety I think is an important way to go, and I think I would certainly endorse that as a way of proceeding.

With respect to the clinical trial design, which is what we have been asked to comment on, some of the literature that we were given to read in advance of this suggested, however, that looking at clinical markers such as pulmonary function, statistically, based on -- I would agree with what Dr. Szymanski said.

Depending on when the initial treatment is begun, we may or may not see that much in terms of change in pulmonary function or, worse, continue to see deterioration, but there was some suggestion from these studies that pulmonary function and the study thereof would take 500 or so subjects in a prospective randomized trial, whereas the marker of the CT scan would certainly say less.

Again, that was certainly strongly stated in the Dirksen paper, and the point made in some of the other supporting literature.

I would just be a bit concerned about FDA's requirement to have industry show an effect on pulmonary function as a way of conducting post-marketing surveillance with respect to efficacy as opposed to the CT scan model, not knowing too much about this disease, just based on what we were given to read.

MS. BAKER: I would also agree with some of the prior comments about collaboration, that the patient groups have so eloquently stated.

Given that this a rare disorder, there must be a finite number of centers that have developed the expertise and a decent number of patients that they care for.

With the increasing costs for clinical trials, three separate IRBs as opposed to one, hiring data managers, hiring research nurses, I know there are a lot of political implications, but it would be more cost effective to the clinicians also if there was more collaboration in the design and the implementation of studies. So, I would recommend that.

DR. QUINN: Just a couple of comments to follow up on that. One would be having the common protocol go from the FDA to the three different companies, and it sounds like that is what is being posed to us. Is there a common protocol that would be acceptable to the companies, to the FDA and so forth?

The second is, besides myself, if there is anyone else here from the NIH, what about NHLBI? Are they involved in research on this disease, and are they interested in supporting a clinical trial on this, again, coming back to help the companies.

This is a rare disease. It is going to need federal support. If these products are clinically efficacious, the numbers are just so small we will never get to it, unless there is a massive undertaking to reach out to as many as possible.

The third issue is, in the design of the study, the best is going to be a placebo controlled trial, randomized.

The horse is out of the barn. Everyone is using these products, it sounds like, with established disease, or a lot are. Is it going to be difficult to get people to be randomized into a placebo arm, to really get you your best data.

I raise that. I don't have an answer for it, but it is the best clinical design. Maybe someone can address that who has thought about doing these trials.

DR. ALLEN: I thoroughly agree with your statement about the need for NIH involvement. It was something that was very apparent to me.

Also, normally, we have as a sitting member a person from the blood component of NHLBI. I would hope that the FDA would assure that the lung division at NHLBI is given a transcript of this committee hearing, and perhaps invited to participate further in the discussions of study design, and perhaps supporting it.

I agree with you, this really does need a coordinated approach and it really, it seems to me, is going to need a long-term approach where we are looking at very early diagnosis before onset of any clinically apparent lung damage, and consideration of, at what point early on might therapy be appropriate, in an attempt to prevent early damage, rather than mid-point.

DR. SZYMANSKI: A very simple question. What is the gene frequency?

DR. DI BISCEGLIE: One in 50 for Z among Caucasians.

DR. SZYMANSKI: One in 50,000?

DR. DI BISCEGLIE: One in 50.

DR. SZYMANSKI: That is not so very rare.

DR. BRANTLY: One in 100.

DR. SZYMANSKI: And the homozygosity frequency?

DR. BRANTLY: One in 2,500.

DR. ALLEN: Please use the microphone, because we are recording. Do you want to stand up and make a brief clarification of the statements you just made?

DR. BRANTLY; I just wanted to go back to the frequency. So, the frequency of the Z allele in the Caucasian population, Northern European is somewhere between 50 and 100. In the United States, it is closer to one in 100 individuals, and the frequency of the ZZ allele among the general population is actually about one in 2,500.

However, among the COPD population, approximately one in 10 individuals is either MZ, SZ or ZZ. So, it is a highly enriched population.

DR. MANNO: Back to the idea of a randomized controlled trial, the one that was presented to us from ZLB, one of the inclusion criteria was an FEV-1 of between 35 and 70, which seems to me significant impairment of function.

So, if we are going to do something randomized and controlled, it is clear we need to be able to identify patients to include early on.

DR. NELSON: Is this clinically possible? The time that patients come to see physicians is when they have symptoms. Isn't that true?

I mean, you would have to screen huge populations in order to do it, and it seems to me that that is not readily feasible without a Katrina type of a budget.

DR. SIEGAL: You ought to be able to collect patients quicker if you enlist the foundations that are interested in this disease and their family members, screen those. You could probably find a larger population to look at early.

DR. ALLEN: Absolutely, I think especially given where the human genome project is now. The issue of developing family cohorts and being able to screen much more easily is something that isn't too far down the line.

DR. PIERCE: Just a real quick comment about the screening issue and the early treatment. Certainly that would be the best outcome, would be to prevent the disease in those who are destined to get it.

One of the little complications of that is that, if you do identify family members, for example, of an index case early, and you follow them, there is a certain percentage of people with a severe deficiency who never in their lifetime, even living into old age, develop emphysema. So, that complicates the picture slightly.

DR. SCHREIBER: I think that the design that ZLB put up is a very feasible one, and I think it has a lot of scientific merit.

Using the CT scan to look at lung function seems to be a very reasonable outcome, compared to the FEV-1. When you look over the data that we got, you have to have a tremendous rate of functionality decline in the FEV to be able to see any kind of statistical difference, something to the tune of 50 percent per year.

If we had 50 percent per year, there wouldn't be any women around like the lady that addressed us. So, that is very, very rapidly advancing disease, and I think what we see is that there seems to be anecdotal evidence, evidence by the ability of people with the condition to have function for a good number of years, that we do have an effective treatment.

Now what we are looking at is our clinical trial to justify that there is an efficacy of the treatment. One of the things that I think would be good is if the companies are designing clinical trials, and they are going to do something like an end point of CT, at least they design them, not that they have to be combined, but at least that you could look at them with some kind of combined analysis, a meta analysis.

Two is too small for meta analysis, but at least you can compare some kind of an outcome based on that type of functionality.

The same we see is that the other end points are very hard to get to, unless we have huge sample sizes, probably more than the patients available, and it would be good to enroll early on patients, because then I think that you might be able to see the change in the functionality at an earlier stage, which would give us more evidence that we would be happy with.

The other European data bordered on significance. If it had just crept over a couple of tenths, a couple of hundredths of a p value, we probably wouldn't be sitting here now, except to discuss whether the three compounds were similar or not, and whether the side effects were similar.

DR. DI MICHELE: I certainly agree, and maybe, Dr. Schreiber, you can comment on this further with your experience in clinical trial design.

The only issue with the ZLB study is, I did not see, certainly in the presentation, what the criteria are going to be for efficacy.

I understand the trial design, and what the primary end point was, but I didn't think I saw the exact measure of rate of increase in lung density or whatever it was going to be that they were going to use as a measure of efficacy, or just slight improvement in one group over the other.

I think the issue is that, because this is a plasma derived protein, and because it is a plasma derivative, it does have certain safety considerations, as was mentioned before, in terms of viral safety issues, and emerging pathogens, et cetera.

I think it is important for this population to actually prove efficacy. I think given, however, the clinical heterogeneity, it sounds like in the clinical course, and the actual clinical phenotype and the clinical progression, how much of a parameter change will it take to give this population or the drugs enough benefit of the doubt to say, yes, it looks like there is some efficacy here.

I think, from a statistical standpoint, I don't exactly know how rigorous the trial design is going to have to be, I think, to prove efficacy.

It sounds like, with few patients, clinical heterogeneity, I don't now. I think even the estimate of 130 in the Dirksen paper, you may actually require more than that. I don't know if you had any comments on that.

DR. SCHREIBER: I wasn't clear what their end point was. I thought it was a 15 percent shift, which seems to be a very reasonable end point marker.

DR. GIRGSDIES: May I comment? You are right. We are not knowing yet what is really clinically meaningful, which loss of line density, but we are thinking that we can create these data, and this is just the pilot study, and we are committed to it, and we will also have, after 50 patients, we will have an interim analysis to look whether we should include more than 100 patients. You are right. We are creating data right now.

The other comment I would like to give to you, it is really hard to find patients for a placebo controlled trial where products are on the market.

I had a hard time, I still have a hard time to look all over the world to get centers where no product is approved. So, it is in Australia, it is in Ireland, it is in South Africa. So, it is a hard time for me.

DR. ALLEN: I would suggest that probably, in this day and age and where we are with these products, that a placebo controlled trial is not ethical.

There need to be alternative ways of looking at it, whether it be comparison of different products, comparison of different dosages, or other comparative issues in there. It is almost 1:00 o'clock. We need to quickly address B and C. Dr. Nelson?

DR. NELSON: I just wondered how the clinical trials are going to deal with important co-factors like cigarette smoking in this population.

I don't know. It is at least additive. It may be synergistic. It certainly complicates it. The genetic defect is not the only thing that causes emphysema.

DR. ALLEN: And certainly those confounding factors need to be looked at very, very carefully. Comments on the adequacy of the proposed safety monitoring program or other suggested actions for the FDA?

DR. KUEHNERT: I think it would call for reinstituting the registry that NIH had had. I think you can get an awful lot of really good information about treatment and side effects and everything else from the registry.

NHLBI just funded a MARFANS registry, and I would think that a condition like this, a registry is an excellent way to go to answer many of these questions that we are looking at, particularly given that the centers that would enroll people in a federally funded research registry would have the capabilities to do some of these more esoteric testings, like CAT scans.

It certainly would have made the adverse event data that was reported much more meaningful, had we had some context like that to weight it against, and to be able to track back to individual patients.

DR. KLEIN: It seems to me that the discussions about efficacy are a different study set than about safety, particularly on an ongoing basis.

I mean, the way this came to attention is that something changed in the product, and who knows what is going to change down the road.

So, you do a six-month efficacy study, fine, you look at efficacy, but that does not address the long-term safety issues.

I don't know what the solution. I saw this nice handout from the Alpha-1 Foundation, which I thought was very well thought out, but I am wondering, how is this going to be supported, who is going to pay for this.

These are the same issues that, frankly, exist for other plasma products. I don't have an answer, but I think it is something that we need to think about in a broader way, because these needs are here, not only for these patients, but for others as well.

I mean, I think there needs to be some lowest common denominator for surveillance for safety, and I just think that Medwatch isn't enough.

DR. ALLEN: I certainly agree with you on that. I think some of the proposed ways to address it was outlined. I thought it was a very good first start. Other comments or questions?

DR. KLEIN: Just two comments. A registry would serve the kind of purpose you have in mind, or do you think that would be insufficient?

DR. KUEHNERT: I guess I don't know enough about it to know exactly how things get reported, and is this in any sense of real time, or is it reviewed every year? How often do people look at the data?

DR. KLEIN: I guess you were talking about long-term effects of chronic treatment or chronic disease. I guess those things could be worked out, but you certainly wouldn't need something real time to look at that.

DR. KUEHNERT: No, but I mean, probably more often than once a year or something, but it wouldn't need to be daily but maybe once a month or something.

DR. KLEIN: Finally, I would just like to go on record, if I may, I rarely disagree with our esteemed chairman, but I think that maybe in the United States it would be extraordinarily difficult to do a randomized trial with placebo, but in many areas of the world where the drugs are not available, it is not all that difficult to do, although I sympathize with your problems of recruitment.

I think it is absolutely crucial. Here we have a treatment, an intravenous treatment for life, where we are really not convinced that we are doing anything to help patients.

I think that, before you commit human beings to that, you really need to know that you have some evidence that you are going to do them some good.

So, in areas where it won't be quite as difficult, perhaps, to recruit subjects, where drug perhaps is not available, I would encourage the company to do so.

DR. DI MICHELE: With respect to safety, I agree with a lot of what has been said. There may need to be sort of more long-term comprehensive collection of data in the registry format, and I wouldn't disagree with that.

I think that the important issues of safety, to a certain extent, are going to be tied in a little to efficacy.

I mean, should the efficacy trials, for instance, show deterioration in pulmonary function if they continue to do that, or certainly an acceleration of CT changes with product. Obviously, that may end up being a safety issue in and of itself.

The other issues, though, that have been brought up are impact of what has been presented to us today on bioavailability, and for instance, immunogenicity.

I think there probably does need to be some element of design in safety studies that look at immunogenicity and that look at bioavailability in terms of whatever can be monitored in terms of pulmonary function.

That could be a subset of a larger group of patients that are just followed prospectively on a more passive basis in terms of passive reporting, but prospectively with respect to adverse events in general, where issues can be picked up and then looked at in a more organized fashion in terms of potential etiology.

DR. CRYER: I just mention that the beauty of having a system where the three trials have similar end points in time in their design would make it so that only one company had to do a placebo group, because the others could compare to that one. It sounds like one already has to committed to it. So, I think that is crucial.

DR. EPSTEIN: I wonder if I could ask a technical question of the representative from ZLB. Dr. Pierce pointed out to us that, In the Dirksen study, alpha-1 PI was better than placebo on CT with a trend of significance of a p of .07, but under a different analysis of the CT, p was only equal to .2.

That seems to raise a fundamental methodological question about whether we know the correct way to interpret the CT data. I just wonder if I could get a comment on that from the ZLB representative.

DR. GREGSDIES: I am not a statistician. So, I can't really answer your question. Sorry for that.

DR. ALLEN: Other important comments? Dr. Epstein, have you received the information that you need? Do you have specific points you would like to have clarified by the committee?

DR. EPSTEIN: Yes, I think we have had a very good discussion by the committee, and that a lot of the comments will prove very useful to us.

DR. SCHWARZ: I would like to clarify something, because it came up by one of the committee members, how products are released, controlled, lot to lot.

Clearly, there are a lot of specifications which must be met, and one of those is specific activity and the elastic activity, of course.

DR. DI BISCEGLIE: I am sorry, I had to leave the room for a minute, but I did want to make a comment about safety monitoring related to liver disease. As a hepatologist, I would be remiss in leaving that out.

We know that this product would not be expected to have any efficacy in the treatment of the liver disease, but I wonder if there is any evidence that it might make the liver disease worse.

The liver injuries caused by accumulation of the protein within hepatocytes, is it possible that if you raise the serum levels, you inhibit the secretion of the protein from hepatocytes, and might exacerbate underlying liver disease that occurs. So, I would encourage monitoring of liver function.

DR. ALLEN: I think that is a very important statement. Thank you.

DR. DOPPELT: I just wanted to add one point .I think everybody -- rightfully -- is focusing on the efficacy, but if the FEV-I is a difficult parameter to measure, and theoretically you would like to get people into the studies earlier, but that may not happen, and you are not quite sure how to evaluate CTs -- there may be different methodologies -- then the number that you need to get an answer may be unreachable.

DR. EPSTEIN: I may have spoken a little bit too quickly. It has been pointed out that we would like some discussion about the issue of communications at this point in time, either through modifications of product labeling or a dear doctor letter, as has been suggested by the Alpha-1 Foundation or Association, I am not sure which.

DR. DI MICHELE: I agree. In fact, I was just about to say that, that I think that, with respect to the issues that were brought up, certainly safety surveillance, which was the first issue that was brought up here today, I think you could get a jump start on certainly voluntary reporting, by making sure that both patients and their doctors, and potentially even ERs, have this information and, you know, really ask for diligence in reporting. I would agree with certainly the dear doctor letter, as suggested by the Alpha-1 Association.

DR. SCHREIBER: I guess I have to just take the opposite. I saw nothing at all that convinced me that the small shifts on electrophoresis constitutes any health effect, and the data that we saw on the health effects was not convincing that these new products were any different than the one on the market. So, I would not put out a dear doctor letter at all. I think that there is just no evidence to go on.

DR. DI MICHELE: I guess what I would say is, it doesn't need to be an alarming letter. I think just a request for information, if that is what was being raised here today.

It sounds like the companies have gone back and done some considerable work to try to look at the reasons for this heterogeneity.

I mean, it certainly seems to be enough of a concern, I mean, within the FDA in terms of product quality, et cetera, that it would seem to me that what you are calling for is just saying, what are you seeing out there in the community.

I am not sure it needs to be an alarmist letter saying we think that there are going to be some problems here. Just, please, you know, examine or be aware of these issues, and that we would like to know if you see anything of any concern, and these are the adverse events that we are interested in.

DR. ALLEN: I certainly would support that position. In particular, given the broader discussions that were presented by FDA staff on biological products, the issues of manufacturing, storage and so on, and the whole issue of heterogeneity, my guess would be that most physicians in practice don't have a significant appreciation of that, and they presume that what is presented is the ideal, when in fact it isn't.

So, I would think a letter that certainly points out some of these issues with a reasonable degree of brevity and clarity for the average practitioner would be important, and it then raises the need for collection of further data, the adverse event reporting system.

Again, I think developing a registry with cohort monitoring, if we can get the NIH to support some of these studies, would be extremely important for long-term follow up.

DR. KLEIN: I guess the question is, what do you put in a dear doctor letter. This is already a hot issue in the community. I don't have to tell any of the alpha-1 people about that.

I have seen letters to the editor submitted that are going to be published. So, they are already aware that there is a potential problem.

We have sat through, as I said, several elegant presentations that suggest to me, at least, one, that there has been no change in the one product from the time of the phase III study.

There has been an observation, but it turns out that it is not a change in the product since the data that were used to license the product.

The second thing that I think I learned is that there is probably no evidence that any of the small changes in the protein have been related to anything that relates to either safety or efficacy of this product.

I suppose, if I were trying to write a dear doctor letter, I guess that is what I would say, that this has been an issue, and right now there is literally nothing that we can say to suggest that there is either a threat to the safety of the patient or a decrease in the potential efficacy of the product, and I don't know what to say beyond that. If there are other ideas, I would like to hear them.

DR. DI MICHELE: Historically, you know, several years ago we had a somewhat similar situation where -- you know, think outside of the box in terms of a dear doctor letter.

We crafted a letter that was informative, in terms of, here are the issues on the table, here, you as the physician, are things that you can do with your patient, and then report back to the company, and then the company can take that information and, in turn, provide it to us in some sort of form that we could, in fact, turn it into something more useful.

DR. KLEIN: I thin the problem, as I see it, is that the patients are being taken care of by probably physicians who are really experts in lung disease, but probably have no biochemical expertise, I might say, than I do.

So, sending them data and western blots is not going to be helpful, unless you can interpret those data for them. I would, again, stress a cautious interpretation from what we have heard today.

DR. DI MICHELE: I agree. It is a little bit of a difficult issue, and that is exactly what I would think that you might be able to send out, is something like, look, this is an issue that we are grappling with. We don't know if it is an issue. We would like you, the care takers in the community, to help us by, number one, thinking about potential adverse events and, number two, reporting them in such a way that they are interpretable.

For instance, you might want to give age and sex and, God forbid, IGA status and a few other things like that, that might help both industry and the FDA really begin to get a handle on exactly -- well, on this issue, and just to put this issue to rest. I don't see anything wrong with that, without causing major alarm in the community.

DR. SNIDER: I think a voluntary reporting system of health effects will not work. There has to be a system that is prospective, that is carefully designed, and preferably is used by the three companies.

That probably can't be physician based, but nurse or IV therapist based, because these people are getting weekly infusions, and with the development of such techniques, I think we would be much further ahead than we are now.

DR. DI MICHELE: If I can just say, I absolutely agree with you, and I think what was just asked was an interim measure while everything is getting started. Otherwise, I couldn't agree with you more.

DR. ALLEN: I am going to call this session closed. I want to thank everybody, all of the temporary and permanent committee members and the guests.

I think Donna and I both have enjoyed working with you over the last two days. Some of you will be leaving the committee and, again, thanks very much, and I look forward to working with you in the future in some other capacity.

[Whereupon, at 1:14 p.m., the meeting was adjourned.]