FOOD AND DRUG ADMINISTRATION

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

9:01 a.m.

DR. WEISS: I'm going to ask everyone to take their seat, please. We'll be starting shortly. I would like to call this meeting of the Ophthalmic Devices Panel to order and note that there is a quorum present. We will have introductory remarks by Sally Thornton.

MS. THORNTON: Good morning. Permit me to introduce myself. I'm Sara Thornton, Executive Secretary for the panel. On behalf of the FDA I would like to welcome you to the 107th meeting of the Ophthalmic Devices Panel.

Before we proceed with today's agenda, I have a few short announcements to make. First of all, I would like to remind everyone to please sign in on the attendance sheet on the registration area just outside the meeting room here. All public handouts for today's meeting are available at the registration table.

Messages for panel members and FDA participants, information, or special needs should be directed through Ms. AnnMarie Williams who is available at the registration area. The phone number for calls to the meeting area is (301) 590-0044.

In consideration of the panel, the sponsor and the Agency, we ask that those of you with cell phones and pagers either turn them off or put them on vibration mode while in this room and to make your calls, please, outside the meeting area. Note the flyers on the door.

Lastly, will all meeting participants please speak directly into the microphone and give your name clearly so that the transcriber will have an accurate recording of your comments.

Now, at this time I would like to announce the voting member appointment of Dr. William Mathers of the Casey Eye Institute in Portland, Oregon. Dr. Mathers has been appointed to serve until October 31st of 2007.

I would like to welcome our Acting Industry Representative, Mr. Andrew Balo, Vice President for Regulatory and Clinical Affairs with DEXCOM, Inc. in San Diego, California. Mr. Balo also serves as the Industry Representative on the Neurological Devices Panel. Mr. Balo is sitting in for our panel Industry Representative Mr. Ronald McCarley who has recused himself from today's panel deliberations.

Will the remaining panel members please introduce themselves beginning with Glenda.

MS. SUCH: Glenda Such, Lighthouse International, Consumer Representative.

MR. BALO: Andy Balo, Industry Representative.

DR. SCHEIN: Oliver Schein, Wilmer Eye Institute, Johns Hopkins.

DR. BANDEEN-ROCHE: Karen Bandeen-Roche, Department of Biostatistics, Johns Hopkins.

DR. McMAHON: Timothy McMahon, Department of Ophthalmology, University of Illinois at Chicago.

DR. BRADLEY: Arthur Bradley, School of Optometry, Indiana University.

DR. MACSAI: Marian Macsai, Evanston Northwestern Healthcare, Northwestern University.

DR. GRIMMETT: Michael Grimmett, the Bascom Palmer Eye Institute, the University of Miami.

DR. WEISS: Jayne Weiss, Kresge Eye Institute, Wayne State University School of Medicine.

DR. MATHERS: Bill Mathers, Oregon Health Sciences University.

DR. CASEY: Richard Casey, Charles Drew University, Jules Stein Eye Institute, Los Angeles.

DR. COLEMAN: Anne Coleman, Jules Stein Eye Institute, UCLA.

DR. VAN METER: Woody Van Meter, the University of Kentucky in Lexington.

DR. HUANG: Andrew Huang, University of Minnesota.

DR. ROSENTHAL: Ralph Rosenthal, Division of Ophthalmic and ENT Devices, FDA.

MS. THORNTON: I'd like to just announce that Dr. Janine Smith who will be in attendance at the panel will be here in a very short time.

I'd like to now read the conflict of interest statement for the meeting on February 5, 2004. The following announcement addresses conflict of interest issues associated with this meeting and is made part of the record to preclude even the appearance of an impropriety.

To determine if any conflict existed, the Agency reviewed the submitted agenda for this meeting and all financial interest reported by the committee participants. The conflict of interest statutes prohibit special government employees from participating in matter that could affect their or their employer's financial interest.

The Agency has determined, however, that the participation of certain members and consultants, the need for whose services outweighs the potential conflict of interest involved, is in the best interest of the government. Therefore, waivers have been granted for Drs. Michael Grimmett, Oliver Schein, and Woodford Van Meter for their interest in firms that could potentially be affected by the panel's recommendations.

Dr. Grimmett's waiver involves an imputed interest, a grant to his institution for the sponsor study in which he has no involvement and is uncompensated. Dr. Oliver Schein's waiver involves two consulting arrangements, one pending for a competitor's unrelated device for which he has not received any compensation, and the second with a competitor's unrelated device for which he receives an annual fee between $10,000 and $50,000. Dr. Van Meter's waiver involves an imputed interest, a stockholding in the parent of a competing technology firm in which the value is greater than $100,000.

The waivers allow these individuals to participate fully in today's deliberations. Copies of these waivers may be obtained from the Agency's Freedom of Information Office, Room 12A15 of the Parklawn Building. We would like to note for the record that the Agency took into consideration other matters regarding Drs. Anne Coleman, Arthur Bradley, Michael Grimmett, Andrew Huang, Marian Macsai, Oliver Schein, and Jayne Weiss.

Each of these panelists reported past or current interest involving firms at issue but in matters that are not related to today's agenda. The Agency has determined, therefore, that the panelists may participate fully in all discussions.

In the event that the discussions involve any other products or firms not already on the agenda for which an FDA participant has a financial interest, the participant should excuse him or herself from such involvement and the exclusion will be noted for the record.

With respect to all other participants we ask in the interest of fairness that all persons making statements or presentations disclose any current or previous financial involvement with any firm whose products they may wish to comment upon.

I would like to now read the appointment to temporary voting status. Pursuant to the authority granted under the Medical Devices Advisory Committee Charter dated October 27, 1990, and as amended August 18, 1999, I appoint the following individuals as voting members of the Ophthalmic Devices Panel for this meeting on February 5/6, 2004.

Karen Bandeen-Roche, Ph.D., Richard Casey, M.D., Marian S. Macsai-Kaplan, M.D., Oliver Schein, M.D., Andrew Huang, M.D., Janine Smith, M.D., Woodward Van Meter, M.D.

For the record, these individuals are special government employees and consultants to this panel or other panels under the Medical Devices Advisory Committee. They have undergone the customary conflict of interest review and have reviewed the material to be considered at this meeting. Signed, David W. Feigal, Jr., M.D., MPH, Director, Center for Devices and Radiological Health. Dated January 20, 2004.

Thank you, Dr. Weiss.

DR. WEISS: Thank you, Sally.

We will now open the open public hearing. I will read a statement which was requested by the FDA.

"Both the Food and Drug Administration and the public believe in a transparent process for information gathering and decision making. To ensure such transparency of the open public hearing session of the Advisory Committee, FDA believes that it is important to understand the context of an individual's presentation.

For this reason, the FDA encourages you, the open public hearing speaker, at the beginning of your written or oral statement to advise the committee of any financial relationship that you may have with the sponsor, its product 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, the FDA encourages you at the beginning of your statement to advise the committee if you do not have such financial relationships. If you choose not to address this issue of financial relationships at the beginning of your statement, it will not preclude you from speaking."

I would call Glenn Hagele to the podium as the first public speaker. You have up to 10 minutes.

MR. HAGELE: I need some assistance with the video. Dr. Weiss, with your permission, could I come after the following speaker?

DR. WEISS: Why don't you stay up there if they can arrange that, Mr. Hagele, because we have a written letter from someone and perhaps we can use this window of time to read the letter while you're preparing your --

MR. HAGELE: Thank you.

DR. WEISS: If that would be agreeable.

Sally Thornton has a letter that was sent in from someone who wanted to participate in the open public hearing but was not able to appear.

MS. THORNTON: This is a letter from Peter D. Van Patten, M.D., the Duluth Clinic Virginia in Virginia, Minnesota.

"Dear Ms. Thornton. I had planned to make a short presentation at today's meeting but could not attend due to a scheduling conflict. If possible I would like to have my following comments read into the record during the appropriate time slot of the meeting.

My name is Peter D. Van Patten. I have practiced ophthalmology since 1991. I am also a subject in the U.S. clinical study of the ARTISAN Myopia Lens and have bilateral ARTISAN implants. I have no financial interest in the ARTISAN lens or Ophtec, the sponsor of this study.

The purpose of my testimony is to provide additional information to the FDA and the FDA panel for consideration during today's discussions. Prior to receiving ARTISAN lenses my refractions were -10.0 X -0.75 in both eyes. Previously I was having increasing problems with contact lens wear to the point where the symptoms became intolerable.

After considering all available options, I decided to proceed with the ARTISAN lens implant. My left eye received the ARTISAN lens in February '99, five years ago, and my right eye received the lens in March 2001, nearly three years ago.

My current refractions are -0.75 X -0.5 in the left eye and plano X -0.5 in the right eye. I have an uncorrected acuity of 20/30 in the left eye correctable to 20/20 and 20/20 uncorrected vision in the right eye correctable to 20/15. My outcomes were very successful and my overall vision is excellent.

I typically wear glasses only for night driving. I have experienced mild night glare on occasion postoperatively that was not present prior to receiving the lenses. However, I would rate the level of glare as minimal.

I have had significant functional improvements during my high visual demand activities such as ophthalmic surgery. Also, I would rate my daytime vision as suburb. I consider both procedures to be a success. Over the past five years I have continued to discuss the ARTISAN lens as an important investigational surgical option with my patients whom I found to be appropriate candidates for open ARTISAN lens clinical trials.

Based on my experience as a subject in this study, it is my opinion that the ARTISAN lens is a safe and effective lens when implants by a skilled surgeon. I would ask that you consider my comments during your discussions and hope that you are able to make a favorable recommendation today so as to make this technology available to others who seek correction for high myopia. Sincerely, Peter D. Van Patten, M.D."

DR. WEISS: Thank you, Sally.

Mr. Hagele, are you ready?

MR. HAGELE: We are coming up momentarily.

DR. WEISS: Sounds good. If you're still having difficulty, I understand that Ms. Woodlock does not have slides so she perhaps could do her presentation while you are getting that ready.

MR. HAGELE: Thank you.

DR. WEISS: Ms. Woodlock. Would you mind, perhaps, giving your presentation from the table instead? Thank you.

MS. WOODLOCK: I am Leslie Woodlock, Patient Advocate of the Surgical Eyes Foundation. We are a nonprofit organization whose constituency is consumers with sub‑optimal outcomes from refractive surgery. Our goals are simply to raise awareness of the risks of elective eye surgery, provide support and identify solutions for patients living with complications, and advocate for informed decision making. I personally became involved with the Surgical Eyes Foundation after failed LASIK surgery in 2000.

I am here today to discuss the safety of phakic lOLs. While much of SEF's concern with the ICL was discussed at this panel's meeting on October 3, 2003, we would like the panel to address the following issues:

Diameter selection is critical for centration of this device since under sizing could result in a failure of the lens to vault the anterior capsule properly, resulting in contact of the device with the capsule and subsequent anterior cortical cataract development.

The need for a tight fit is recognized by the applicant and yet selection of the ICL diameter is to be based on the white‑to‑white measurement. Since no exacting correlation between the white‑to‑white measurement and the ciliary sulcus diameter exists, how will patients be protected from secondary cataract development?

The increasing thickness of the physiologic lens with aging as well as during accommodation means that the desired post‑operative vault of the ICL will fluctuate and actually diminish over time. This has the potential to accelerate the development of anterior cortical cataracts.

The incidence of endothelial cell loss is a repeated concern throughout the previous discussion. In the event that the ICL must be removed following a noted progression of anterior capsular opacities, there is no evidence suggesting that explantation of the ICL will be less harmful to the endothelium than its continued presence.

Further, in cases of either device‑induced or naturally occurring cataracts, the ICL will have to be explanted before the implantation of a pseudophakic IOL. Clearly, for all patients, a second and possibly third intraocular procedure must be entertained with further potential for loss of endothelial cells.

Continuing with our concern for loss of a functional endothelium, the dynamics of a shallow anterior chamber depth and progressive endothelial cell loss is unknown at this time. Most cases of Fuchs' endothelial dystrophy do not become clinically evident until patients are approaching their fifth decade. Will implantation of the ICL result in an even earlier loss of endothelial integrity and, ultimately, penetrating keratoplasty?

These patients would appear to be at even higher risk for endothelial cell loss regardless of an allowable standard for minimal anterior chamber depth of 2.8 or 3 mm. It is not possible to assess risk for younger individuals at the time the ICL is implanted since they will not have visible indications for the condition.

Revisiting the effects of aging of the physiologic lens, another consequence is the shallowing of anterior chamber. The applicant has found a correlation of shallow anterior chamber depth to endothelial cell loss. It is reasonable to suspect that aging of the crystalline lens and subsequent reduction of the anterior chamber depth will put older patients at increased risk for decompensation of their corneas secondary to endothelial dystrophy.

In regard to implantation of this device, typically the risk of cystoid macular edema increases with each intraocular surgical procedure. In the case of device‑induced cataracts with subsequent explantation followed by implantation of a pseudophakic IOL, the potential for CME would be significantly greater.

Correct positioning of the ICL requires a tight sulcus to sulcus fit with anterior displacement of the iris. The fact that the potential narrowing of the anterior chamber angles following implantation was not consistently examined via gonioscopy in the PMA suggests only a cursory concern with the potential for narrow angle glaucoma. Patients with naturally narrow anterior chamber angles as well as those whose angles will narrow subsequent to aging, are at higher risk for development of glaucoma.

The presence of the ICL vaulting above the anterior capsule changes the dynamic of the posterior iris and its contact with the anterior capsule. The potential for pigment dispersion is very real as the ICL haptics rub against the posterior iris.

Pigment dispersion has a known occurrence in the general population but does not manifest until the fifth decade. Implantation of this device in younger patients with a predilection for pigment dispersion will quite conceivably accelerate the process and lead to pigmentary glaucoma.

Anterior cortical cataracts, narrow angle glaucoma, pigmentary glaucoma and endothelial dystrophy are naturally occurring conditions but are potential complications of the ICL. A very real possibility exists that health insurers will not cover the cost of treatment for these conditions since they could be viewed as secondary to an elective procedure.

SEF is already aware of patients receiving corneal transplants following corneal refractive surgery who were denied reimbursement by their health insurer for this very reason. The negative impact on the patient is two fold. Either they will be denied coverage for a naturally occurring medical condition or they will have to pay for the deniable complications secondary to an elective surgery.

The optical diameter of the ICL is listed as 4.65 to 5.5 mm. While the diameter of a posterior chamber ICL cannot be compared to the typical, stated ablation diameters of LASIK and PRK, it is interesting that the optical diameter is so small. Pseudophakic lOLs are typically in the 6.0 mm range and there are still patients who will notice glare and halos under low light conditions.

Using our knowledge of pseudophakics' experience as a guide and, given that the population having ICL surgery would typically be much younger with larger pupils, it would seem very certain that many individuals will experience unwanted glare and haloes from spherical aberrations created by the uncorrected rays of light passing through the peripheral physiologic lens.

Continuing on with the discussion of the optical diameter effects, it is necessary to mention the recent publication of Dr. Steven C. Schallhorn's study suggesting the irrelevance of pupil size to visual quality under mesopic and scotopic light conditions, in particular, that pupil size does not correlate with night driving performance.

This oft touted study, however, does nothing to explain why numerous journal articles by leading refractive surgeons suggest the use of brimonidine tartrate (Alphagan), an adrenergic agonist that suppresses pupil dilation to produce a relative miosis, as well the direct‑acting miotic, pilocarpine, be used post‑operatively to suppress the ill‑effects of night time driving in refractive surgery patients. The Surgical Eyes Foundation bulletin board is overflowing with empirical evidence from our patients and our participating doctors of the effectiveness of pupil constricting agents in the reduction of low light glare and halos. Our bulletin board already has one ICL patient complaining of this very thing and two well‑known refractive surgeons recommended Alphagan as the remedy.

With regard to quality of vision, we ask that PMAs for all forms of vision correction devices be stratified by pupil size. The PDA should mandate that quality of vision be measured objectively with wavefront and other objective tests that have been utilized by optical scientists like Dr. Raymond Applegate that stratify results by pupil size, and that these results be published and made readily available to consumers with regard to any form of vision correction device.

We have had many patients of all ages with large pupils post on our bulletin board about nighttime visual aberrations, regardless of refractive error. We understand that an effective optical zone on the corneal surface and the optic diameter of a lens that sits behind the iris are not comparable; however, we feel very strongly that patients with large pupils are at risk with this device.

One common experience of patients visiting our web site and bulletin board is in regards to the informed consent agreement. While explanations of potential visual and physiological complications are discussed, patients typically do not understand the chronic and irreversible nature of those complications.

Informed consent continues to be a major concern of SEF for elective refractive surgery. Unnatural visual effects seem to impact deeply on many patients sense of well being. The psychological emotional aspects of vision complications are not something potential patients can understand or be prepared to accept following negative outcomes.

This completes my presentation. On behalf of the board of trustees of the Surgical Eyes Foundation and our constituency, I wish to thank the Advisory Panel for the opportunity to express our concerns. Thank you very much.

DR. WEISS: Thank you very much.

And you will be limited to 10 minutes for your presentation.

MR. HAGELE: That should be more than adequate. Good morning and thank you for the opportunity to address this panel. My name is Glenn Hagele. I am the Executive Director and founder of the Council for Refractive Surgery Quality Assurance, which from this point forward I will refer to by its acronym CRSQA.

In the way of disclaimer, I have no financial interest in AMO or the ARTISAN phakic intraocular lens. My travel here today is self‑funded. Although I am the Executive Director of CRSQA, the opinions I express are my own and do not necessarily represent the opinions of individuals affiliated with CRSQA.

CRSQA is a nonprofit consumer/patient organization that through its sister websites USAeyes.org and ComplicatedEyes.org receives over 800,000 visitors annually. We provide objective information about refractive surgery issues and resources for those unfortunate few who have encountered a poor refractive surgery outcome. Additionally, CRSQA evaluates and certifies refractive surgeons based upon patient outcomes.

In addition to research of published studies and case reports, my interaction with patients provides me with a unique accumulation of anecdotal information and a perspective of a patient. The issues and concerns I will raise today all relate to communication between physician and patient.

Potential refractive surgery patients, especially high myopes and high hyperopes, seek options. With a greater understanding of the advantages and limitations of corneal‑based refractive surgery, those with high refractive errors find the probability of achieving the convenience of a reduction of the need for corrective lenses less than spectacular.

The phakic intraocular lens has been available outside the United States for the better part of a decade, and it is reassuring that this panel will have the opportunity to determine if a new option will be available to Americans.

Not surprisingly, I have some concerns. I will leave to others to debate clinical data, and raise only those issue that from a patient perspective are of equal importance.

Pupil Size. Capt. Steven Schallhorn, MD of the United States Navy recently presented a significant performance‑based task study of 105 consecutive LASIK subjects to determine what effect preoperative scotopic pupil size has on postoperative night vision.

Dr. Schallhorn's, and subsequent studies, found no direct correlation between scotopic pupil size and reaction‑based visual task performance. Although Dr. Schallhorn's study may

provide evidence that pupil size alone is a poor predictor of induced night vision problems, I have never heard Dr. Schallhorn say pupil size is not important.

Pupil size may be a poor predictor of night vision problems, but as any doctor who has prescribed pilocarpine or Alphagan can attest, pupil size is the moderator of night vision problems, when they exist. Although these are two very separate issues, I ask that this panel be mindful of their interrelation.

Furthermore, the corneal‑based LASIK procedure is not an intraocular lens. Even further, it is not a phakic intraocular lens. Decades of intraocular lens development have shown the importance of edge design and pupil size in regard to halos, starbursts, and glare in low illumination environments. It seems unreasonable to disregard this body of knowledge, regardless of the conclusions of Dr. Schallhorn's findings.

Should this panel ultimately decide to approve the device presented today, I respectfully ask the panel to consider including in the labeling for both physician and patient that the probability of induced night vision problems when the scotopic pupil is larger than the size of the full optical correction of the device is not easily determined.

I respectfully ask that the patient labeling include a representation of these effects and explanation of probable limitations on the patient, including difficulty driving at night and reading in low illumination environments.

Learning Curve. Today you will have the advantage of evaluating the safety and efficacy of the proposed device when care is provided by what can only be described as some of the best surgeons in the world. I submit that if this device is approved, it will be utilized by doctors who are, shall we say, of somewhat lesser distinction.

With reports of as much as 20% incidence of anterior sub‑capsular opacities with the first few patients of other intraocular lenses when implanted by novice surgeons, it appears self evident that proper implantation of an phakic intraocular lens requires not only training, but practical experience.

I have no reason to doubt that the Sponsor will provide significant training in this regard, and I have equally no doubt that this panel will insist on adequate training and proctoring. I believe, however, it is in the best interest of the patient to be informed of the experience of the prospective surgeon.

Our organization provides a list of 50 Tough Questions For Your Doctor for patients to use as a guide in selecting a refractive surgeon. In our 50 Tough Questions we recommend that a patient seek a doctor who has performed at least 100 refractive procedures of the exact type intend to use on the patient, with the same equipment, and the same refractive error, and significantly more practical experience with similar surgical techniques.

While this panel may find our recommendation of 100 a bit conservative and even restrictive, it does seem reasonable to assume the patient would like to know if he or she is the doctor's first unsupervised phakic intraocular lens patient.

I respectfully request that this panel include in the patient labeling a statement indicating that training and practical experience of the surgeon may be an important factor in the probability of a desirable outcome.

Induced Intraocular Pressure. This panel is much better qualified to determine the safety of the Sponsor's phakic intraocular lens than I, but it appears reasonable to assume that the patient will require periodic evaluation of intraocular pressure during use of the phakic intraocular lens. Who will pay for this care?

Phakic intraocular lens for the convenience of a reduced need for corrective lenses is an elective, arguably cosmetic, procedure. The patient who is making the decision to proceed is making this decision partly based upon cost.

If significantly elevated long‑term care were required to maintain good ocular health after phakic intraocular lens implantation, the probable costs for examinations, visual fields, and medication to manage a surgery‑induced chronic condition would most probably be an important factor in the patient's decision to elect to have surgery in the first place.

I doubt that it is within the power of this panel to require a doctor to provide long‑term cost estimates preoperatively, but it does seem reasonable that the patient labeling include an indication of the type and frequency of reasonably probable surgery related long‑term care.

I'm sure that when presented with this probable treatment plan, the patient will not need the labeling to recognize that these costs should be a part of the decision regarding the relative value of a reduced need for corrective lenses.

Endothelium. There seems to be a lack of clear consensus on the long‑term effects of phakic intraocular lens on the quantity and quality of endothelium cells. In the clinical trials, a mandatory evaluation regime underlies the importance of this consideration. The Sponsor is requesting approval for implantation in patients as young as their twenties. Assuming that the phakic intraocular lens would be utilized until natural cataract development when a person is in his or her sixties, the functional life of a phakic intraocular lens may be as much as 40 years. During this time, the need for regular evaluation of endothelial cell loss seems obvious. Again, who is going to pay for these costs?

Like long‑term care for induced intraocular pressure, it seems reasonable that the patient labeling include some indication of the type and frequency of reasonably probable surgery related long‑term care.

Summary. The issues I raise all relate directly to the communication between doctor and patient. All suggestions are for the purpose of promoting that communication. If properly informed of the immediate and long‑term issues relating to the Sponsor's phakic intraocular lens, I believe that those patients who elect to have phakic intraocular lens implants will have reasonable expectations and will be able to make the decision that best

meets their needs and desires.

Lastly, I do hope that during the course of discussions today I will not hear the term "implantable contact lens". If this is a contact lens, then I've been wearing explantable phakic intraocular

lenses when I water ski. Thank you very much for your time.

DR. WEISS: Thank you. I have been told that there is someone in the audience who wanted to also participate in the open public hearing.

DR. JOHN: Yes.

DR. WEISS: Okay. You have, well, Dr. Grimmett said eight minutes but actually it's now down to seven. If you could identify yourself and any potential conflict.

DR. JOHN: Yes, ma'am. Hi. I'm Maurice John. I'm an ophthalmologist from Louisville, Kentucky/Jeffersonville, Indiana, all in the same metropolitan area. I'm medical monitor for Ophtec. I am not paid by them at all except they paid for my plane fare and my hotel today.

I have no stock which is very good news for Ophtec in that I don't have stock in their company. They would be in trouble. I implanted intraocular lenses starting in 1975. I did radial keratotomy in 1980. In 1993 I had a laser in Sao Paolo, Brazil and we believe the first LASIK was performed with my laser by a colleague of mine in 1993. In 1995 I started doing LASIK in Sao Paolo to get ready for the United States.

In October of 97 I was fortunate enough to implant the first five ARTISAN lenses in the United States. Prior to that I had gone to Brazil and that summer of '97 implanted a couple of lenses down there. Now I've done 200 plus ARTISAN lenses, the majority of which have been myopic and about 10 percent hyperopic.

Starting out in October '97 I found out that there certainly is a learning curve to implanting this lens which has already been mentioned. It is a short but steep learning curve and there is an advantage to being a good surgeon.

Mr. Hagele's excellent presentation mentioned that he encourages his patients to ask for a surgeon who has done 100 or more cases and that's going to be very, very difficult with an ARTISAN lens because there just aren't many of those people out on the planet. I have a large, busy, refractive surgery practice and I don't know what that number should be but I've been doing it six years and, like I say, I've just done 200 plus of those.

This lens needs space to be put in the eye, there's no doubt about that, but there is adequate technology to make those measurements to determine if there is adequate space. I would also like to comment on glare. Having done radial keratotomy since 1980 I can assure you that all those patients had starburst and glare and that did not kill radial keratotomy.

Then I've done between five and 10 patients who are in the subset of people who have larger than 6 mm pupils and none of them have glare. I strongly think the reason for that, especially in this population of people who are between -10 and -20 primarily they've had glare, super glare, all their life. So if they get glare from this, it's pretty much peanut glare and then if it's a killer, then this lens can be removed really quite easily.

After that point the problems are primarily if you estimate the anterior chamber depth they are surgeon related and we've seen that time and time again. I introduced this lens in Brazil, as I said, in October of 1997 to my friend Eduardo Martinez who we think is the first guy to do LASIK in North or South America. He was using other phakic IOLs and has switched to this and now gives paper presentations on it.

I have been to South Africa many times. I go to a meeting over there every two years and I introduced it in 1998 to some of my colleagues there. They also had access to all the phakic IOLs that are available throughout the world.

My colleague, Jan Venter, is up in England now and he is working for a consortium and he gets referred all of the anterior segment surgeries that these LASIK boutiques find. In September of last year he implanted 100 of these lenses. That's how much he believes in their efficacy.

The nice thing about this lens, having done a lot of refractive surgery, when I'm in the office and seeing patients, I walk by and I pull the chart off, I look at it and I see it's an ARTISAN patient and I am so happy because I know that I'm going to be in and out of there quickly and that these patients are going to see well and we have not beat up their cornea trying to do -10.0 or 12.0 diopters on them.

If they see 20/30 they are far happier than a 20/25 LASIK patient. It's amazing. My LASIK patients are always whining. They have some slippage, especially the -7.0, -8.0, -9.0, -10.0 and they are always wanting enhancements even though they are 20/25. These ARTISAN patients have tremendous quality of vision.

It's amazing to me. I just keep reminding myself you're taking the very worse people on the plant, the one or two percent, bottom or top percent, depending on how you want to look at it, and basically pretty much nailing them, knocking a homerun every time up to the plate.

My feeling is that patients should run to this lens and I've had some patients who you say FDA study and you've got to wait three months between eyes and they've gone elsewhere. I've seen a couple of them come back and they said, "I should have listened. I should have come."

The problem we have is, and I had this in 1996, people wanted tried and true RK. They didn't want LASIK. We had the same thing here where 98 percent of these people's friends had LASIK, you know, quick, fast, next day, the American way, and this is a bit of a journey. There are some people who have not had it and it's so unfortunate. I think this lens is wonderful and thank you very much. I hope I beat my seven minutes.

DR. WEISS: By 60 seconds. Thank you.

We will now close the open public hearing and we are going to move on to the open committee session starting with the division update. Dr. Rosenthal.

DR. ROSENTHAL: Thank you, Dr. Weiss. First, let me say that I very much appreciate Donna Lochner coming today because she is theoretically no longer with our division. She has taken a detail with the Division of Cardiovascular Devices as their Deputy Director but has come back to deal with this lens today. We want to wish her all the best of luck on her detail and thank her again for all the hard work she's done for our division and I know she will do a lot of hard work for the Division of Cardiovascular Devices.

Secondly, just as a reminder to all companies, and I'll say this tomorrow as well, but it is important that all companies who are dealing with PMAs with our division schedule a pre-PMA meeting to discuss accountability, stability, safety and efficacy even if they have submitted numerous previous PMAs or PMA supplements. This will help ensure better submission and one that will be less likely to result in a nonfiling decision or result in significant measure deficiencies.

I make these comments because the MDUFMA goals will have to be met in 2005 and the quality of this submission will help us considerably should it be excellent to meet our review goals. Thank you.

DR. WEISS: Thank you. we will now have branch updates by Donna Lochner and Everette Beers.

MS. LOCHNER: Thank you. I am pleased to announce to the panel that Morcher's PMA P010059 was approved by FDA on October 23, 2003. This PMA was for the endocapsular tension ring which is used for capsular bag stabilization in patients with pseudoexfoliation syndrome or other situations of compromised zonules. You may recall that this PMA was reviewed by the panel in January of 2002.

I'm also pleased to announce that Eyeonics' PMA, formerly C&C Vision, P030002 was approved by FDA on November 14, 2003. This PMA was reviewed by the panel in May of 2003. The PMA was for the CrystaLens Accommodating IOL which is intended for primary implantation in the capsular bag for the visual correction of aphakia in adult patients in whom a cataractous lens has been removed and is intended to provide near, intermediate, and distance vision without spectacles. The CrystaLens provides approximately 1 diopter of monocular accommodation.

Thank you. That concludes my announcements.

DR. WEISS: Thanks, Donna.

DR. BEERS: I'm Everette Beers, Chief of the Diagnostic and Surgical Devices Branch. Since our last update in May of 2003 we have approved three PMAs, P020050 for the WaveLight Allegretto Laser for Myopia and Astigmatism, Ms. Jan Callaway, team leader. The approved indication was for a LASIK correction of myopia up to -12 diopters with or without astigmatism up to -6 diopters.

We also approved P030008 which, again, was a Wavelight Allegretto laser for Hyperopia and Astigmatism. Let me back up. The WaveLight Myopia was approved October 7, 2003. This one for WaveLight Allegretto for Hyperopia was approved October 10, 2003.

Again, Ms. Jan Callaway was the team leader.

The approved indication for this WaveLight Allegretto Laser was for LASIK correction of Hyperopia up to +6.00 diopters sphere with up to +5 diopters of cylinder with MRSE up to +6 diopters.

On October 10, 2003, we approved P990027/S6 for the Bausch & Lomb Zyoptix, Ms. Daryl Kaufman team leader. The approved indication here was for Wavefront-guided LASIK correction of myopia up to -7 diopters with up to -3 diopters of astigmatism and with MRSE of up to -7.5 diopters.

We've had no staff changes since the last update in October. During 2003 we cleared approximately 36 510(k)s. This concludes my update.

DR. WEISS: Thank you very much, Everette. That will conclude the branch updates. I just wanted to say for the panel, Donna, that we've all valued all the hard work and the great work you've done and we're going to miss you. Good luck in your new position.

I will now ask the sponsor to come to the podium for presentation of PMA P030028. There will be one hour for the presentation. Each presenter should speak into the mike, identify yourself and your relationship with the sponsor and any potential conflicts.

MR. McCARLEY: Good morning. I'm Rick McCarley, the President and CEO of OPHTEC USA which is based on Boca Raton, Florida. OPHTEC USA is a wholly owned subsidiary of OPHTEC BV based on Groningen, the Netherlands. We are the sponsor of the PMA under review today for the ARTISAN Myopia Lens.

First, I would like to thank the panel for their time in preparing for today's meeting, especially the primary reviewers for their indepth review and comments. I would also like to thank the FDA team of Dr. Lepri, Dr. Toy, Dr. Gray, and Dr. Lu for their extraordinary effort during the last six months bringing this PMA to panel.

Finally, I would like to thank the audience for their interest and presence at today's meeting to observe the review of the ARTISAN lens. Today's presentation will be made by Dr. Vance Thompson, an ophthalmologist from Sioux Falls, South Dakota, and Dr. Doyle Stulting, Professor of Ophthalmology at Emory University, Atlanta, Georgia.

Dr. Thompson is an investigator in the Artisan lens study but holds no financial interest in the ARTISAN lens or OPHTEC. OPHTEC did pay for Dr. Thompson's travel expenses today.

Dr. Stulting was an investigator in the ARTISAN study and was engaged by OPHTEC following the PMA filing as a consultant. He and Maurice John of Jeffersonville, Indiana/Louisville, Kentucky are medical monitors for this study.

Also today with us is Dr. Camille Budo from Belgium. Dr. Budo is a medical monitor for the ARTISAN lens studies in Europe and is a paid consultant for OPHTEC BV. He will be available during the day to answer questions related to the ARTISAN lens usage outside the United States.

Finally, Dr. Stan Bentow, the statistician for the PMA, is here to assist as needed. Dr. Bentow is the Department Manager of Biostatistics and Data Management for Advanced Medical Optics. OPHTEC has a business relationship with Advanced Medical Optics for the worldwide distribution of the ARTISAN lens.

With that said, I'll turn the presentation over to Dr. Thompson.

DR. THOMPSON: I'm Dr. Vance Thompson from Sioux Falls, South Dakota. I do not have a financial interest in the ARTISAN lens and my expenses for being here today are being covered by OPHTEC.

It is a sincere honor of mine to present my experience with the ARTISAN Phakic Intraocular Lens implant to the FDA Ophthalmic Devices Panel. After completing a fellowship in refractive surgery with Dr. Dan Durrie in 1990 I entered into private practice in my home state of South Dakota.

I've been integrally involved in multiple FDA monitored clinical trials as a primary investigator in United States eximer, PTK, PRK, and LASIK clinical trials. I have completed 20 FDA monitored laser and implant clinical trials at my center.

When I was first asked to be a part of the ARTISAN trial, I actually respectfully declined. In 1997 I had a hard time imaging that we would be putting an implant in the eye to correct refractive error. I received a call from an international investigator that I respect who shared with me his positive experience with this implant and asked me to look into this further.

As a result of this call, I chose to go to the Netherlands and study with the inventor of the ARTISAN lens, Professor Jan Worst, to find out more for myself about this lens. I was impressed with it's long track record. I hadn't been real familiar with it at that point.

I basically came away with the feeling that the lens itself had some unique safety features that explain its excellent performance internationally and with good surgical techniques the outcomes were outstanding.

I saw patients who had had the lens implanted 12 years previously, five years previously, one week post-op, one day post-op. I had a real good experience there and I was impressed enough to then accept the invitation to be a part of the United States clinical trial.

So I came back home and implanted my first ARTISAN lens in September of 1998. I have 74 eyes included in the data being presented today. I was surprised at how quickly I became comfortable implanting this lens and how quiet these eyes looked postoperatively.

A hundred percent of my patients have bee pleased with their outcome and they have all opted to have their fellow eye performed. With continuing enrollment I have performed a total now of 95 ARTISAN lens implants. After having used it, I can't imagine not providing this quality option for my patients in my practice.

This is a single piece PMMA lens for the correction of myopia. It is elliptical in shape and 8.5 mm in length. It comes in two optic diameters of 5.0 and 6.0 mm. It has a slight anterior vault to create a safety zone between it and the crystalline lens.

The ARTISAN lens is designed for implantation into the anterior chamber of the phakic eye. It is fixated to the mid-peripheral iris by incorporation of a portion of the anterior iris stroma into an opening in the haptic with an instrument specifically designed for this purpose. This process of lens fixation is known as enclavation.

Here is a post-mortem specimen from an 86-year-old who died from an unrelated cause after implantation of an ARTISAN lens six years previously. Note how quiet and undisrupted the posterior uveal pigment appears.

This slit lamp photograph shows an example of the appropriate amount of Iris tissue that should be incorporated into the lens haptic for stable lens fixation. Proper fixation requires incorporation of about 1 mm of iris tissue between the aligned arms of the haptic. Keep this picture in mind because we are going to be showing you a photograph of an inadequately fixated lens later in this presentation.

An advantage of this lens is the ease with which one can attach it to the iris or detect it from the iris. It can be repositioned during surgery for optimal centration or it can be easily removed by pushing the iris tissue back through the opening in the haptic.

The lens is available in two optic zone sizes, a 6 mm optic in powers of -5 to -15 diopters and a 5 mm optic in powers of -5 to -20 diopters. Here is a Scheimpflug photo showing the healthy separation of the intraocular lens from the cornea and the crystalline lens. This is an 18 diopter lens in a patient with an anterior chamber depth of 3.4 mm.

Since the lens attaches to a relatively immobile peripheral portion of the iris, the pupil can be dilated nicely for an unimpeded view of the retina. The lens is implants through 5.2 or 6.2 mm incision and fixated by incorporating a portion of the mid-peripheral iris into an opening in the haptics with the enclavation instrument.

The surgery is performed utilizing a cohesive highmolecular weight viscoelastic. A peripheral iridotomomy or iridectomy is required to avoid pupillary block.

Here is an edited video of one of my patients who had an ARTISAN implant and the first step is marking the limbus for the surgical incision and then marks are placed approximately 10 mm apart to locate incision site for the enclavation instrument. Then the initial vertical limbal incision is made and then dissected into clear cornea.

The entry sites for the enclavation instrument are then created, first here on the right and then now on the left. Viscoelastic is instilled with care to avoid overfilling the anterior chamber while providing protection for the corneal endothelium and also the crystalline lens. The anterior chamber is then entered.

If necessary, more viscoelastic can be instilled. The ARTISAN lens is then rinsed. Then it's implanted with Budo forceps, forceps that help to stabilize the lens. The lens is positioned over the pupil. I like to start with the lens slightly inferior to the pupil since it tends to move superior during lens fixation.

The lens is then enclavated first on the right making sure to incorporate the proper amount of iris tissue, at least 1 mm in width. Here we can see how easy it is to incorporate additional tissue to assure adequate fixation. More viscoelastic can be used to maintain a nice comfortable space between the lens and the endothelium. The left haptic is then enclavated.

At this point in time I would like to put in balance salt solution and then the wound is closed partially. Before the last suture is tied, the remaining viscoelastic is removed from the anterior chamber. Then the last suture tied.

Early designs of this iris fixated lens have been implanted for 25 years with more than 400,000 lenses implanted in aphakic eyes to date. In 1986 a second design known as the Worst-Fechner lens was introduced for implantation into phakic eyes. However, there was concern that the increased vault of this lens might not provide sufficient clearance from the corneal endothelium.

In 1991 the lens was refined to address these concerns. Note the new profile. This design known as the ARTISAN lens has been used successfully worldwide since 1991. The aphakic iris fixated lens is used all around the world and it is also frequently used as a secondary implant particularly during penetrating keratoplasty.

Stable fixation over time has been shown with normal long-term pupillary function and no iris atrophy. With iris angiogram studies, in this case two months postoperatively, it's been shown that there's no vessel disruption or leakage and normal pupillary function is maintained.

The current ARTISAN lens design has been used for 13 years with over 100,000 myopic, hyperopic, and toric lenses implanted worldwide by more than 5,000 physicians to date. The ARTISAN lens is the most commonly implanted phakic intraocular lens in the world today. It is the lens of choice accounting for almost two-thirds of implants in markets where multiple phakic IOLs are available.

I'd like to now review the results of European multi-center study of 518 eyes implanted from 1991 to 1999 at nine sites with -5.0 to -20 diopters of myopia utilizing the 5 mm ARTISAN phakic IOL. Three-year follow-up on 249 eyes has been reported in the literature by Budo and coauthors.

Best spectacle corrected visually acuity was better than or equal to 20/40 in 93.9 percent of patients. Uncorrected visual acuity was 20/40 or better in 76.8 percent of patients regardless of postoperative goal. 57.1 percent were within 0.5 diopters of their target refraction and 78.8 percent were within 1 diopter of their target.

Mean endothelial density changes in a subset of 129 eyes were as follows. After six months there was 4.8 percent cell loss; 6 months to one year, 2.4 percent; one year to two years, 1.7 percent; two years to three years 0.7 percent. Notice the relatively low amount of cell loss and stabilization over time.

The results of the European multi-center study demonstrate refractive stability and good predictability. They concluded there was a favorable risk benefit ratio and that efficacy and safety through three years was demonstrated in this study. There are no published reports indicating long-term safety concerns with the current lens design.

Corneal decompensation and glaucoma have not been reported. International experience with complications and secondary surgical intervention parallels that in the U.S. clinical trials. I have a lot of confidence in this lens design which has been in use since 1991. My patients and I both appreciate the fact that it is removable and exchangeable. My personal experience has been excellent.

I find comfort in the long-term performance demonstrated in the European study and the published literature. The two-thirds market share for phakic IOLs means a lot to me when the majority of surgeons in the world who have their choice of which phakic IOL to implant choose the ARTISAN lens.

I consider this a quality surgical option for high myopes and I also consider this a quality surgical option for low myopes who are not good candidates for other refractive procedures. I am enthusiastic about the results from this lens and I look forward to its approval. Thank you very much for your attention.

I would like to now turn the podium over to Dr. Doyle Stulting who will review the results of the United States clinical trial.

DR. STULTING: Good morning members of the Ophthalmic Devices Panel and the FDA. I'm Doyle Stulting, Professor of Ophthalmology at Emory University. I'm one of the medical monitors of the ARTISAN Phase III clinical study and a paid consultant to OPHTEC. It will be my pleasure to present the results of the U.S. clinical investigation of the ARTISAN myopia lens for the correction of high myopia.

This was an open-label, noncomparative study of patients with 4.6 to 22 diopters of myopia. The lens was available in one diopter power increments and two physical designs. The 5 mm optic was available in powers from -5 to -20 diopters and the 6 mm optic was available in powers from -5 to -15. There were eight postoperative visits.

Note, however, that the study was initially planned to spend two years and later extended to three years. The results were obtained from all implanted lenses by all investigators. Specifically, investigators for this study did not receive training outside of the U.S. or experience outside of the U.S. before they began to participate in the clinical trial.

To be included in the original study subjects had to be 21 to 50 years old with a stable manifest refraction or refractive cylinder of 2 diopters or less, an anterior chamber depth of 3.2 mm or more, and an endothelial cell density of 2,000 or more per millimeter square. The low-light pupil size had to be 4.5 mm or less because only the 5 mm optic was available at the time of study initiation. There can be no ocular disease or abnormality that would affect safety.

The FDA granted a number of protocol waivers to allow implantation in patients who did not meet all of these criteria until the original protocol was amended with expansion inclusion criteria in November of 2000.

These expanded inclusion criteria allowed enrollment of eyes with clinically insignificant and stable peripheral lens opacities, astigmatism of 2.5 diopters or more, anterior chamber depths of less than 3.2 mm, age over 50, pupil size greater than the size of the optic, best spectacle corrected acuity of less than 20/40, and implantation of lenses that would not completely correct the refractive error.

Outcome measures included uncorrected visual acuity, best spectacle corrected acuity manifest in psychoplegic refraction, contrast sensitivity, intraocular pressure, endothelial cell density, and slit lamp observations.

At the time the protocol was developed, cataract formation was not identified as a significant risk because of a six-year history of implantation internationally without cataract induction, and the position of the implant well clear of the crystalline lens. Thus, the approved protocol required only clinical grading of cataracts rather than standardized grading of lens opacities.

In 1997 when the study was initiated a variety of instrumentation was permitted for obtaining specular images and only one image was required for each eye at each visit. As the available technologies and the knowledge advanced, the sponsor changed investigational procedures consistent with FDA, ANSI and ISO discussions in developing guidelines.

This led to a recommendation from the sponsor that sites use only the Konan non-contact specular microscope and obtain three satisfactory images for analysis for each eye at each visit.

Six hundred and 84 subjects were enrolled. There were 662 subjects in the primary study 478 of which were implanted bilaterally. Twenty-two were implanted for compassionate use. Enrollment is ongoing.

The PMA defines several groups for analysis. For this presentation, most safety analyses were based on all implanted eyes while efficacy studies were based on first eyes. Accountability was adequate and the study is ongoing. Two hundred and 32 first eyes have completed three years of follow up and 357 subjects continue to be followed.

This PMA filing is based on 386 eyes which were followed for three years. At three years 62.4 percent of eligible eyes have completed their exams with a large portion of the remainder still to be seen. In judging these numbers the sponsor feels that it is important to be mindful of the fact that this was originally designed and powered as a two-year study. Subjects were recruited with this understanding and some of them elected not to return for their three-year visit.

The number of discontinued subjects was low with only eight percent lost to follow-up. Demographics were not unusual for refractive surgery population with a mean age of 39.6 years. The mean preoperative spherical equivalent refractive error was significantly higher than the mean error of patients seeking refractive surgery in this country today. It was -12.3 diopters. The range was 4.6 to 21.9 diopters.

As we move forward it is important to remember that many of the subjects in this study are high myopes who have no other alternatives for refractive surgery. This population is also at increased risk for undesirable outcomes such as cataracts and retinal detachment. The mean lens power was -12.6 diopters ranging from 5 to 20 diopters.

Let's look at the safety of the lens. One hundred percent of first eyes has best spectacle corrected acuities of 20/40 or better at two and three years after surgery. As you can see from this slide, best spectacle corrected acuity was better at one, two, and three years postoperatively than it was preoperatively.

At three years 49 percent of eyes gained best spectacle corrected acuity while only 6.2 percent lost best spectacle corrected acuity. Two eyes lost two lines of best spectacle corrected acuity. No pathology was reported in either of these eyes so the loss is believed to be due to measurement variability.

I want to pause with this slide because it shows something that previously approved forms of refractive surgery do not, improvement in best spectacle corrected acuity after surgery. Indeed, the fact that only two eyes in this study lost vision is a remarkable result in view of the degree of preoperative myopia and the age of the subject population.

The incision size --

(Whereupon, off the record from 10:14 a.m. to 10:19 a.m.)

DR. STULTING: -- possible dislocation. There were 20 of these. The majority of these procedures were performed at a single investigational site. Most of the procedures were preventative measures taken to avoid possible dislocation. The sponsor advocates comprehensive training of surgeons to minimize similar events after approval.

This graph shows the number of secondary surgical events as a function of the number of implants performed by the investigators in the clinical trial. It is clear that the majority of secondary surgical interventions occurred during the early surgical experience.

Approximately 50 percent of the events occurred among the first 10 subjects implanted. A disproportionate number occurred at one site and the majority was due to improper lens fixation. These procedures include both preventable and therapeutic interventions.

These data show that there is a learning curve for some of the skills required for successful implantation of the ARTISAN lens. Retinal detachment occurred in six eyes during the observation period. This represents an incidence of 0.3 percent per eye per year in a population of eyes with myopia between 11.5 and 18.6 diopters. This rate is comparable to reported rates of retinal detachments in the literature in high myopes.

This slide shows the incidence of lens opacities in the study. Most were not visually significant or lens related. Only one eye lost two lines of vision to 20/30. Most lens opacities were nuclear which would not be expected to be related to an implanted lens. Only a very few were anterior or subcapsular opacities which would be expected if they were intraocular lens related trauma to the crystalline lens.

This slide summarizes significant lens opacities requiring cataract extraction during the study. One occurred after removal of viscoelastic and intraocular lens implant for high intraocular pressure.

A second was in a 50-year-old with a family history of cataracts. The third was in a 56-year-old man with a preoperative lens opacity and a family history of cataracts. The sponsor does not believe that the incidence of cataract in the study population is unexpected given the age and refractive error of the study cohort.

These are the visual outcomes in eyes that underwent secondary surgical intervention. Even in this group more eyes gained best spectacle corrected acuity than lost best spectacle corrected acuity compared to the preoperative value. The 3.7 percent represents two eyes that lost two or more lines of best spectacle corrected acuity.

Here are the details pertaining to these two eyes. One was due to a retinal detachment and a subsequent macular hole 20/70. The other was due to posterior capsule or haze following cataract extraction and intraocular lens implantation. After YAG capsulotomy this eye had a best corrected acuity of 20/30.

The Agency requested data on adverse events that have occurred since submission of the PMA. There were two of these. One was a cataract extraction that was necessitated by repair of a retinal detachment with trauma to the crystalline lens.

The second was reattachment of a lens that was dislocated during a boxing match. I'm not sure whether he won or lost. The most recent visual acuities in these two eyes were 20/30 and 20/40. Let us discuss endothelial cell density in greater detail. At the time of initiation of the study in 1977 the protocol allowed a variety of instrumentation and required acquisition of only one image per eye per visit.

As technology improved, the protocol was changed. However, it was impossible to acquire old data according to the improved protocol retrospectively. The data presented to the FDA are consistent with the guidance provided to industry by the Agency and the Ophthalmic Devices Panel at the time the study was designed.

This slide shows the results of endothelial cell counts in the original PMA. Although there was not significant reduction in endothelial cell density, the standard deviations of the measurements were relatively large ranging from 17 to 25 percent.

The data set that was reported in the original PMA was derived from one to three images per eye per visit using a variety of instrumentation. The images were analyzed by various site personnel. Because of the variability there was not good statistical power to rule out significant changes in endothelial cell density.

Review of the raw data led to the conclusion that analysis of the specular images could be improved. The sponsor elected to recount available high quality images after consulting with FDA and experts in the field. Data from 12 sites were chosen because they used the Konan specular microscopes. This instrument is now the accepted standard for the most accurate determination of endothelial cell density.

One reading center was employed for consistency. Only the best quality image was analyzed per eye per visit. There were a total of 353 eyes of 215 subjects producing a consistent cohort of 57 subjects with data at all time points throughout three years. The average number of cells analyzed per image was 109.

Here we see the mean endothelial cell density in all recounted subjects and the consistent cohort. Both showed a slight reduction which would be expected even in the absence of intraocular lens implantation.

The equivalent yearly rate of cell loss ranged from 0.72 percent to 1.59 percent throughout the study when all recounted eyes were analyzed. Note the reduction in the standard deviations in the recount analysis.

In the consistent cohort yearly endothelial cell density loss rates range from 0.71 percent to 1.27 percent. This slide shows percent change for each observation period. Changes between consecutive periods are not statistically different. Similar results were obtained in the consistent cohort.

One site exhibited an endothelial cell loss for the two to three-year period of 4.95 percent which was significantly lower than any of the other sites. The sponsor was recently notified that the site had staffing changes, problems with calibration of the specular microscope, and that the microscope required servicing during this period. The results from this site may not be poolable.

Removing this site from the analysis decreases the loss for the two to three-year period from 2.37 to 1.68 percent. The accuracy and precision of specular microscopy with the Konan noncontact specular microscope is documented in the publication by Nichols and coworkers in which 25 normal subjects were examined on two occasions by two examiners. The mean instrument error was 1.7 percent with a confidence interval from -13 to +16 percent.

The important point to be learned from this published paper is that a cell loss reading of 10 percent is within the confidence interval of measurement error and 13 percent of eyes would be expected to have a 10 percent change or more even if no real change existed.

Here we see the yearly change in endothelial cell density and the two to three-year periodic rate. These are not significantly different from guidance. The average cell loss over time was about 50 cells per millimeter square per year or 1.72 percent per year. The projected mean cell count is about 1,300 30 years after implantation.

There was no change in the percent of hexagonal cells and the coefficient of variation after surgery. These data support the conclusion that the implanted lens does not stress the endothelium because a reduction in hexagonality and an increase in coefficient of variation are hallmarks of chronic endothelial stress such as we see with long-term contact lens wearers.

When the endothelial cell density was analyzed, there was no consistent statistically significant association with gender, age, lens model, anterior chamber depth, or preoperative spherical equivalent manifest refraction.

We conclude that the endothelial cell density loss after implantation of the ARTISAN lens is within the acceptable range. There are no statistically significant differences in loss rates between consecutive periods. The loss that was recorded is within measurement error.

In summary, the ARTISAN lens has a superb safety profile with excellent best corrected acuity. Most secondary surgical procedures were due to inadequate lens fixation and could be prevented in the future by surgeon training and proper attention to surgical techniques.

Even when secondary surgical intervention was necessary, best spectacle corrected acuity was maintained. Lens opacities were generally mild, not visually significant, and unrelated to the intraocular lens. The endothelial cell loss was within the acceptable range.

Let's take a look at efficacy. Ninety-two percent of first eyes targeted for emmetropia with preoperative best spectacle corrected acuities of 20/20 had 20/40 or better visual acuities at three years postoperatively. Fifty percent of these eyes had 20/20 or better visual acuity postoperatively.

This slide gives the details of eyes with uncorrected acuities less than 20/40 at three years after surgery. Most or attributable to residual refractive error, usually residual astigmatism. The sponsor believes that the reported uncorrected acuity of 20/70 in one subject was due to a testing or reported error since this subject had a best spectacle corrected acuity of 20/20 and a minimal refractive error.

Remember that only one diopter lens power increments were available in the study. Subjects were included with more than 2.5 diopters of astigmatism without astigmatic correction. We expect uncorrected acuity to increase after approval because of the availability of half diopter power increments and the use of astigmatic corrective procedures when necessary.

As we improve refractive surgical techniques, the comparison of best postoperative uncorrected acuity to preoperative best spectacle corrected acuity is becoming a more discriminating outcome measure. In this study a remarkable 51 percent of eyes targeted for emmetropia had a postoperative uncorrected acuity better than or equal to the preoperative best spectacle corrected acuity.

71.7 percent of eyes were within a half a diopter the target refraction and 94.7 percent were within one diopter of target. Postoperative refractions were remarkable stable with less than a 10th of a diopter change between six months and one year and between two and three years.

The vast majority of subjects were pleased with the quality of their vision, satisfied with their outcomes, and would recommend the procedure to their friends. Visual aberration such as glare, starbursts, and halos were noted in about the same number of subjects after surgery as before surgery.

The majority of subjects had no change in their reported visual symptoms after surgery compared to preoperatively. The only reported symptom that was more frequent postoperatively than preoperatively was halos. We may see the reason for that in subsequent slides.

We look for correlation between visual symptoms and parameters relating to the lens or the eyes in this study. There was no significant correlation of visual symptoms with the relationship between the lens optic and mesopic pupil size, lens power or refractive cylinder except for halos in refractive cylinder which is probably an explanation for the increase in halos postoperatively.

After approval the sponsor believes that postoperative symptoms due to residual refractive error will be reduced because of the availability of half-diopter lens power increments and the use of additional surgical procedures to treat residual astigmatism.

A 20 diopter myope with a 2.5 diopter corneal astigmatism preoperatively who has 20/40 uncorrected acuity after implantation and residual nighttime glare may be ecstatic about his or her surgical result but raise concern among panel members because of the presence of nondebilitating visual symptoms at night postoperatively.

Contrast sensitivity was investigated in a substudy involving 31 subjects. Under photopic conditions without glare, contrast sensitivity was better postoperatively than preoperatively. There were similar results with glare reaching statistical significance at four out of five of the measured points, again with better performance after surgery than before.

Under mesopic conditions without glare contrast sensitivity was the same postoperatively as it was preoperatively. The same results are seen under mesopic conditions with glare.

In conclusion, there was no decrease in contrast sensitivity after implantation of the phakic ARTISAN lens. Statistically significant differences where present usually show better contrast sensitivity postoperatively than preoperatively, very different than currently approved refractive surgical procedures.

In summary, the ARTISAN lens offers excellent uncorrected visual acuity, excellent predictability, good stability of refraction, contrast sensitivity that is unchanged or improved, and high subjective satisfaction rates. The sponsor proposes that the ARTISAN lens be labeled for the correction of myopia with lenses from between five and 20 diopters. Although preoperative refractive cylinder greater than 2.5 diopters was an exclusion criterion for the study, the sponsor does not believe that it should be a contraindication for the use of the ARTISAN lens after approval because residual astigmatism can be managed by the placement of the surgical incision site and other techniques.

We suggest that the lens optic size be greater than the mesopic pupil size when possible. However, we note that no correlation was found between the disparity between optic size and pupil size and visual symptoms at night. Because subjects in the study with preoperative pathologies did not have different results than those without pathology, the sponsor proposes that preoperative pathologies not necessarily preclude the use of the ARTISAN lens.

Endothelial cell density minimums for each age group would be acceptable as a precautionary measure at the discretion of the panel and the FDA. Our experience with inadequate iris fixation primarily at one site emphasizes the need for appropriate physician training in the use of this lens.

The sponsor proposes that the lens be made available only to surgeons who have undergone appropriate training including didactic instruction, supervised wet lab training, observation of live surgery, and supervised initial procedures.

There are a number of benefits of the ARTISAN lens compared to other refractive surgical techniques. The ARTISAN provides excellent refractive outcomes. As opposed to other commonly used refractive surgical techniques, the ARTISAN lens leaves contrast sensitivity unchanged or improved.

There is a good safety profile with few complications most of which can be avoided by adequate training and surgical technique and attention to detail during the surgical procedure. Endothelial cell loss was within the expected range and there was very high patient satisfaction.

The lens is exchangeable and removable with good outcomes. It avoids the potential complications of corneal surgery such as scarring, complications of flap preparation, and irregular astigmatism. It provides an effective treatment for myopes, especially those who are not candidates for other refractive procedures.

The sponsor asks that the ARTISAN phakic IOL be recommended for approval. Thank you.

DR. WEISS: Seeing that concludes the sponsor's presentation, I would like to thank the sponsor for the clear presentation and we are going to break for 10 minutes. I would request that everyone be back here promptly in 10-minutes time.

(Whereupon, at 10:40 a.m. off the record until 10:56 a.m.)

DR. WEISS: We are going to begin with questions from the panel to the sponsor so I will ask the sponsor if they could take a seat up front. Okay. I'm going to be changing the format a little bit for the edification of the panel today and what I'm going to be doing is going around the table and asking you what questions you might have for sponsor in the attempt to maximize our time.

I have one question. You mentioned that there were three cases where the pupil size was larger than the optic size. Did those patients have any halo or glare or visual symptoms associated with that?

MR. McCARLEY: Yes, they did and, in fact, there were three patients who had their lenses removed that had the optic size larger than the pupil -- sorry,, the pupil larger than the optic size. But there were many more patients in the study that had larger pupils than the optic.

DR. WEISS: I think I'm confused then. I heard from Dr. Stulting that he identified particularly three patients that had pupil size larger than optic. But from what I'm hearing right now, there were more than three patients?

MR. McCARLEY: That's correct. There were three patients. The slide identifies three patients who had lens removal or secondary surgical procedures as a result of that.

DR. WEISS: Okay. So then there were three patients with lens removal because the pupil size was larger than the optic size and they were symptomatic.

MR. McCARLEY: That's correct.

DR. WEISS: But how many patients had pupil size larger than optic size?

MR. McCARLEY: I'll have to pull the PMA.

DR. WEISS: So you can get that?

MR. McCARLEY: Yes, we can get that.

DR. WEISS: The other question on that, because of glare symptoms at night in relation to one of the people who gave a comment at the open public hearing, were any of the patients needed to be on Alphagan at night?

MR. McCARLEY: Not that we're of specifically for that purpose, no.

DR. WEISS: Okay. So we're going to go around. Glenda, did you have any comments or questions?

MS. SUCH: Two questions small in nature. One is what was the youngest age you actually had in the study group?

MR. McCARLEY: It was 21, I believe.

MS. SUCH: That's what I thought I heard. I can't remember the second question so I guess that's it.

DR. WEISS: Well, we can get back to you.

Mr. Balo.

MR. BALO: I don't have any questions.

DR. WEISS: Dr. Schein.

DR. SCHEIN: I have comments coming up later but only one question now. I'm wondering about data from other sources that could be brought to bear?

MS. THORNTON: Oliver would you --

DR. SCHEIN: My mike is too far away. I'm interested in data from other sources on the device that might be useful. What I've heard so far relates to this long-term series of 19 patients in Europe. It's a consistent cohort but it's a very, very small group.

Dr. Stulting mentioned a publication out of Europe recently but that group reported only 50 percent of the patients that they started with. Are there any data sources available with 100, 200, 300 patients with three, four, five-year follow-up that we can examine?

MR. McCARLEY: Not that we are aware of. Obviously there have been recent publications that have started to come out where more people became involved, especially in Europe in the mid-'90s who have longer term data now.

Dr. Mihai Pop from Canada also has some data that I believe will be produced in the next month or two in one of the major journals. But as far as answering your question, I don't think anyone has done a large study of endothelial cell count for a long term.

DR. SCHEIN: I needn't even be restricted to endothelial cell count. Something would disconnect if there are 100,000 implants that have been done and there is essentially no data externally with high levels of follow-up.

DR. STULTING: This is Doyle Stulting. I can address that a little bit. The European study was 518 eyes with a three-year follow-up on 249 eyes. That's Dr. Budo's paper. I think the number that you were referring to is the endothelial cell density study and that was 129 eyes followed for three years. There probably aren't any rigorously followed series of eyes out there in the literature giving us two or three-year follow-up with the accountability that we would like to see because of the nature of the refractive surgery population. What we do know is the number of implants that have been used and the lack of reports of significant long-term complications so that gives me at least a little bit of comfort knowing that there are so many lenses out there in eyes and, yet, there aren't reports of problems with these lenses long term.

DR. SCHEIN: In Europe is there mandatory reporting with explantation?

MR. McCARLEY: Yes. In fact, it's required as part of the CE process. In Europe they do require you to report adverse events -- all adverse events.

DR. ROSENTHAL: May I just add that as part of the PMA process it's the sponsor's obligation to submit all data that's been published that they know of in the literature and not in the literature about the device. The FDA does receive everything that is supposed to be -- that is out there. It's supposed to be submitted with the application.

DR. WEISS: Dr. Macsai has a question on that point.

DR. MACSAI: I have a question for Dr. Rosenthal. Does that include CE data? Do you share with CE and does CE share with the FDA?

DR. ROSENTHAL: The data we are supposed to receive is the data that the company submits that they know is in the public domain. CE data may not be in the public domain. Many of the countries within the European community consider much of the adverse event data to be confidential.

I think with Britain we do have some sort of mutual agreement that when there are serious problems with the device, we are notified and, likewise, they are notified when we have serious problems. But generally there is not a worldwide sharing of data relating to post-market problems with devices.

DR. WEISS: We're going to go on to Dr. Bandeen-Roche.

DR. BANDEEN-ROCHE: Thank you. I just have a couple of questions about the endothelial cell count data that was presented in the PMA. The first is that there were 12 sites that contributed data to the final analysis. I'm wondering if you can tell us whether you've analyzed data to determine how those sites compared to the sites that did not contribute data to that analysis other than not having the Konan microscope, things like case mix, provider experience, anything like that?

MR. McCARLEY: The Konan machine provides a relatively good image that is standard. And they also provide a separate software that actually provides a way to read images in the standardized way. It was out choice based upon the recommendations of the experts, some of which have testified in front of this panel, which machine is likely to give you consistent good readings. Not favorable readings but to be able to read it at all. In fact, we utilized exactly the same machines as Dr. Edelhauser, for instance, and some of the others around the country that actually do endothelial cell counts.

DR. BANDEEN-ROCHE: But what I'm getting at is that only 10 of the U.S. sites use that microscope and how might those sites have been different than the others?

DR. STULTING: Maybe I could ask Dr. Bentow for a little bit of help. Did you look at the baseline for the two sites?

DR. WEISS: You can identify yourself.

DR. STULTING: For the groups of sites that were included in the endothelial cell versus those that were not.

DR. BENTOW: Yes, this is Stan Bentow with AMO. We didn't look at a comparison with the previous data set because we went with only the Konan pictures that we used in the latter one, although we did look at site comparison and analysis for that data set, the recounted data set.

DR. SCHEIN: Did you make a comparison --

DR. WEISS: Dr. Schein, can you identify yourself?

DR. SCHEIN: Right. This is Oliver Schein. Did you make comparisons within those centers that were using the Konan scope as to who is in versus who is out? By that, I mean you have images on a very small proportion of the total images that were generated even at the sites that used that technology. It would be useful to know rates of adverse events in versus out, baseline cell counts, age, etc. Do you have those to show us?

DR. BENTOW: We can look at that and see if we can bring that up.

DR. WEISS: Thank you. Dr. Bandeen-Roche, if you have no other questions.

DR. BANDEEN-ROCHE: I have one more question that actually goes to that. I believe I read in the updated part of the PMA that images with the number of cells counted less than 70 per reevaluated and if it was felt that they weren't of good quality, they were not eliminated. This seems potentially biasing to me. It seems like that could well undercut the rate of loss by excluding the images with the low cell counts. I wanted to give you a chance to respond in case I'm just not understanding.

DR. STULTING: I understand what you're asking. I can tell you that the site that we mentioned that had the high rate of secondary surgical procedures and what not was one of the sites that was included in the recount data.

DR. WEISS: Dr. McMahon.

DR. McMAHON: Tim McMahon. I have two questions. The first continues along the line of endothelial cell counts. Were test and retest analysis done by individual sites and were there any variances in the 95 percent confidence intervals amongst those sites?

DR. STULTING: No, looking at test/retest for a single site was not part of the protocol. Once again, the protocol was developed back in 1997 when these kinds of questions were really not at the top of our minds and we weren't looking for a technology that gave us the ability to discriminate a .6 percent loss from a 1.6 percent loss over a period of a year. Those parts of protocol that we might want to design today for a very scientifically rigid investigation were not part of the investigation that we are reporting today.

DR. WEISS: We do not have the amount of time to have the sponsor coming up to the podium at this point so we are going to have to continue along with these questions.

Dr. Bradley. Oh, I'm sorry. Dr. McMahon.

DR. McMAHON: This is my second question and it's in a completely different area. I was troubled by the number and percentage of protocol deviations. I think it was as high as over 20 percent and I'm kind of concerned as to what the rationale for that was. There are a couple of cases identified.

In particular, initially three comments or comments about three patients with pupils larger than the optic and now new statements saying that there's more than that. There's protocol instructions and inclusion/exclusion that prohibits that and what is the justification for all these?

DR. STULTING: We tried to address this question in the presentation since it was raised in some of the comments from the panel that were forwarded to the sponsor. The original protocol that was designed in 1997 had exclusion and inclusion criteria.

For example, for astigmatic error that was present before surgery patients who were high myopes who had no other choice of refractive procedures requested ARTISAN implantation. Their surgeons requested it from the sponsor. The sponsor requested protocol deviations for the FDA for use compassionately in these patients and it was granted.

As the time went by eventually in the year of 2000 the FDA and the sponsor expanded the protocol inclusion criteria so that patients with anterior chamber depths less than 3.2 mm, pupil sizes greater than the optic size and high astigmatism could be implanted with informed consent and so these patients were later included in the protocol and that's how they got in there.

It wasn't because investigators enroll people that they shouldn't enroll. It was because the indications were expanded with informed consent, knowledge of the Agency, and a decision on the part of the sponsor.

DR. McMAHON: So the FDA okayed each one of these?

MR. McCARLEY: This is Rick McCarley. Initially we received very few requests for protocol deviations in our study. As time progressed and we believe as surgeons became more comfortable with the procedure, the surgical technique itself, they started to receive more patients and see more patients that they believe could be assisted by it but, in fact, at the same would not be compromised.

We started to get more and more requests for protocol deviations. We worked with the FDA, the Agency, to create an arm, a substudy. It's called Protocol Deviation Substudy. All of the criteria except a certain list of items that Dr. Stulting mentioned were included. These patients signed an additional informed consent on top of the normal informed consent for the study.

DR. WEISS: I would just follow through and I'm just repeating what you said. What percentage of the protocol deviations were granted with the update approval and what percent were not?

MR. McCARLEY: All of them were --

DR. WEISS: A hundred percent.

MR. McCARLEY: A hundred percent were known by the FDA or approved by the FDA. We either gained approval from the FDA on a one-by-one basis before the substudy started and after the substudy started they were approved by the institutional review boards. The patient included them in the normal enrollment.

DR. WEISS: So 100 percent were approved by the FDA before they had the surgery.

MR. McCARLEY: Correct.

DR. WEISS: Dr. Bradley.

DR. BRADLEY: The sponsor emphasized that the procedure led to improved visual acuity and improved contrasensitivity. I wondered if the sponsor had evaluated the relative importance of image magnification and image quality on these changes in acuity in contrasensitivity?

DR. STULTING: We recognize image magnification and we recognize the potential improvement in image quality because of the placement of the corrective lens but nothing was designed in the protocol to look at these things specifically, objectively, and scientifically other than the collection of data that you have in front of you.

DR. WEISS: Dr. Macsai.

DR. MACSAI: I just want to first follow up on your comment, Dr. Stulting. This increased visual acuity that was shown on the slides during your presentation is accountable due to the magnification of the IOL. Correct?

DR. STULTING: Some of it is. That's correct.

DR. MACSAI: And in your contrast sensitivity studies, how was the contrast sensitivity measured preoperatively as for point of comparison? Was it measured in spectacles?

DR. STULTING: Yes.

DR. MACSAI: So what did you expect if someone is -12 in their spectacles that their contrast sensitivity would be decreased versus that with an intraocular lens?

DR. STULTING: Are you asking --

DR. MACSAI: Would you expect these results from placement of an intraocular lens?

DR. STULTING: I think I would and I think I would be pleased.

DR. MACSAI: It's because of the difference between the spectacle and the movement of the lens inside the eye and the lack of changes in refractive index.

DR. STULTING: That's probably correct.

DR. MACSAI: Okay.

DR. WEISS: Dr. Grimmett.

DR. GRIMMETT: Dr. Michael Grimmett. My first question was already addressed by Dr. Macsai and Bradley regarding magnification. The second point I wanted to make was that Dr. Stulting showed a slide regarding the lack of change and hexagonality and coefficient of variation. This was a fairly young cohort of patients, I think, ranging from 21 to 50 something with an average range in the 30s, I believe. As Dr. Edelhauser confirmed on our October meeting when I asked him the same question, a young cohort can have a very robust endothelium and really mask the morphemetric data so it's stress factors that we all think of regarding these changes which may manifest in an older subgroup can be completely hidden in populations this young. I just wanted to point out that fact when we considered the endothelial data. That's all I have.

DR. WEISS: I had one other question in terms of trauma dislocating this lens. Would you then advise patients who were in careers such as boxing or hockey or whatever that that would be a contraindication to having the lens? Basketball depending on how you play it?

DR. STULTING: I think that's a reasonable suggestion.

DR. WEISS: Dr. Mathers.

DR. MATHERS: Dr. Bill Mathers. I seem to feel a sort of disparity between the study -- the mean of the study group and that which the sponsor is asking permission to include later. The mean here was 39 years of age and -12 refraction. It might be that this is actually the group that you feel that this lens is the most advantageous for and has the greatest impact.

In fact, one could say that Dr. Stulting's comment that there is no other choice for some of these people may be a factor but you are requesting permission for patients down to 20 and a refractive error that is much lower than this. How do you -- help me with feeling how these two groups actually compare and the justification for using it in a larger group.

DR. STULTING: The profile of the refractive -- of the patient population in this study pretty much parallels the clinical practices and the refractive populations that are part of publications for refractive procedure in the literature so far. The mean age, in fact, for all of these is pretty consistent at 39.

As a consumer of this technology, I would like to have it available to me to use in circumstances for I feel it is appropriate. There may be a relatively young patient who has a relatively thin cornea who would be an appropriate candidate for it based on parameters other than age.

I would like to have it in my armamentarium so that I can offer it to that person. I think that the selection of procedures for refractive surgery has to be based on many more things than the refractive error and the age.

DR. WEISS: Dr. Casey.

DR. CASEY: Richard Casey. Dr. Thompson, you showed a slide and you made the comment that there was no iris vessel disfunction or leakage in patients in this study, but the title of the slide was an angiogram two months post-op with an aphakic IOL. My question is was there a systematic evaluation by angiogram of patients in this study?

DR. THOMPSON: No. We didn't do angiography in this study. We were basically showing that because some of the disadvantages of other implants in the past than chronic cell and inflair from vessel leakage and we wanted to show the integrity of the blood/aqueous barrier for this lens.

DR. CASEY: My question was related to there was a small number of minority patients in this study and we know that there are patients that have different -- the iris is of different thickness and different vascular density and so issues of inflammation could be important in different subpopulations. If it wasn't done, it wasn't done.

My second question is can you tell us anything about the endothelial cell loss in those patients who required a second surgical procedure and were they followed after the second procedure to determine if there was any accelerated rate of attrition of endothelial cells?

DR. STULTING: We can try to get those numbers specifically for you after lunch but I can reiterate the point that I made before and that is that the site that had most of the secondary surgical procedures was one of the sites that was in the recounted endothelium cohort so we have a good bit of information on those patients in particular. Let me make a note of that and we'll try to get the data. The specific question is endothelial cell counts on people with secondary procedures.

DR. CASEY: Yes.

DR. WEISS: Dr. Coleman.

DR. COLEMAN: Yes, this is Anne Coleman. I had a question regarding your exclusion criteria for patients with glaucoma. How is that determined for those individuals to be excluded? Was it by visual field and optic nerve evaluation, or was that by clinical judgment?

DR. STULTING: I'm afraid we didn't probably use the strict criteria that a good glaucoma specialist would request. It was based on clinical diagnosis and the use of medications.

DR. COLEMAN: And then --

DR. STULTING: A refractive surgeon's diagnosis, I guess.

DR. COLEMAN: And then at one, two, and three years how many of the patients were on chronic glaucoma medications for maintenance of intraocular pressure?

DR. STULTING: Ten out of 1,147.

DR. COLEMAN: Thank you.

DR. WEISS: Dr. Van Meter.

DR. VAN METER: Woody Van Meter. A couple of questions for Dr. Stulting. Early on in your presentation you were talking about training and mentioned that all investigators were trained for this study in the United States. Is that correct?

DR. STULTING: That's correct.

DR. VAN METER: We had anecdotal evidence from Dr. John about multiple procedures done in South America and South Africa and those patients were not part of this study. Dr. Grimmett has already covered my concerns about the statistical legitimacy of the endothelial data and I guess we can talk about that later.

On slide No. 60 that you showed, there was reference to a 56-year-old with a preexisting cataract who had a family history of cataracts who had an ARTISAN lens implanted. That seems to be a little bit outside the box.

MR. McCARLEY: Rich McCarley. There were actually two patients in the study that had a family history of cataracts but that wasn't found out until after the patient had actually received the implant. In one case I'm very familiar with, the surgeon implanted the first eye and six months later was getting ready to implant the second eye and noticed the cataract in the first eye. Upon further interview with the patient found out, in fact, it was familial.

DR. VAN METER: Well, I'm not concerned about the family history as much as I am the 56-year-old who was already presbyopic and nearing cataract age anyway must have had it noticed beforehand since it was called a preexisting cataract.

MR. McCARLEY: It wasn't and the patient chose the surgery. The surgeon felt that there was a possibility that it wasn't likely to develop.

DR. VAN METER: Okay. Thank you.

Dr. Stulting, slide 80 you mentioned that proper training will reduce the incidence of complications. Since we have data here from the finest surgeons in the world doing these cases, how are you going to alter the training technique that is listed in slide 102 to reduce complications when mere mortals try to do the surgery?

Slide 102 lists a number of objectives from training, most of which I believe are already, as I peripherally understand it, part of the ARTISAN training program. Can you tell me how you are going to change the training?

DR. STULTING: I'm not exactly sure that I understood the question. Could you repeat it?

DR. VAN METER: Yes, sir. On slide 70 you mentioned that proper training will reduce the incidence of complications. Slide 102 you list the training proposal but, as I understand it, this training proposal is pretty much how training has existed for ARTISAN investigators.

DR. STULTING: I don't think -- there is no question that this surgery is different from what ophthalmologists are used to performing as you could see from the video clip. There is bimanual dexterity that is involved. It's a little bit greater than the bimanual dexterity that we are used to having in other procedures that we perform. That will have to be taught.

As a result of the clinical trials, there are techniques that we have learned that need to be taught perhaps differently, emphasized differently than were done in the clinical trial. We think that those will be possible to teach and that, I guess what you said was, mere mortals will be able to perform those techniques.

After all they do in the rest of the world outside the United States using the data that we showed you from Market Scope with implantation of phakic IOLs. This is the most common phakic IOL that is implanted outside of the United States where ordinary surgeons have a choice of intraocular lens implants to use and this is what they choose to use.

We think that the experience that we have had has made us better at picking out skills that we need to teach and in recognizing methodologies that can be taught to improve the performance and that's what we've learned from the clinical trials.

We would prevent all complications from the lens? Probably not. We still see complications from cataract surgery and other procedures that we perform. I don't think technique will be any different but I think that the risks are well worth the benefit.

DR. VAN METER: Thank you.

DR. STULTING: May I pass the mike off to Dr. Thompson?

DR. THOMPSON: Just a quick comment. I consider myself a mere mortal and I get way more stressed out going into cataract surgery than I do going into doing ARTISAN. I have not found the training to be difficult. Approximately after five implants I had a nice comfort level so I do not think we are going to have a hard time getting ophthalmologists comfortable with this procedure.

DR. WEISS: Dr. Smith.

DR. SMITH: Janine Smith. I wanted to echo Dr. Bandeen-Roche's concerns regarding any data you might have on differences in the cases of patients that had gradable specular images for the endothelial cell counts.

I wonder if you have any data on the proportion of eyes that have gradable specular images in the 12 sites that had the Konan microscope? So that's one, the proportion. The second is could you identify any difference between the cases and people that had gradable images and the ones who didn't.

My concern is from my experience with that particular instrument which has issues that I'm sure we'll talk about later, it happens to be the corneas that have some abnormality that it is much more difficult to get good images in so you can understand why this might be an important question.

DR. STULTING: I appreciate the comment and made note of it and we'll try to address it. I can tell you from having looked personally at many of the images that were obtained during the first part of the study the problem with the images wasn't that there were very few cells, large cells, and unusual cells with Gute and other abnormalities. The problem was focusing, properly counting and whatnot. They were technical problems but I've made note of the question.

DR. WEISS: Dr. Huang.

DR. HUANG: I have some concerns about the safety and efficacy for this procedure in the low myope patient. As we all know, there are many refractive surgery options for the patient with low myopia nowadays. I'm just wondering that in the low myope patient with slightly shallow anterior chamber is the safety equity maintained and achieved in the efficacy of this procedure? Is it just as effective or as safe as some other existing procedures?

DR. STULTING: That's a good question and a good consideration. We looked at endothelial cell losses in patients who had more narrow anterior chamber than those who had deeper chambers and didn't find a correlation with that.

Having said that, I appreciate your concern. The sponsor believes that this is a technology that should be made available so that it can be used at the discretion of the well-trained and discriminating refractive surgeon.

With the information in hand and the proper training, that surgeon can make a reasonable decision about what is the best technology to offer the patient. It's possible that a low myope may do better with an intraocular lens implant because of his corneal anatomy.

Perhaps it's someone who has questionable form fruste keratoconus and the surgeon doesn't want to take a chance on getting ectasia postoperatively. In that particular case the balance may fall toward an intraocular lens implant when for the routine patient with low myopia a corneal procedure may be most appropriate.

DR. WEISS: Thank you. We have one question from Dr. Macsai and we have one from Ms. Such and we'll do those two and then we'll conclude this portion.

Dr. Macsai.

DR. MACSAI: On the panel someone had asked about the endothelial cell counts in the patients that had their implant repositioned, etc. I was wondering if you could give us the endothelial cell data on the entire Group E because I did not have that to review prior to this meeting and I would like to see the consistent cohort within Group E, the entire Group E. I think that is incredibly important because, as Dr. Van Meter stressed, we are mere mortals and what happened in that group is important.

DR. STULTING: I'll add my name to the list of mere mortal ordinary surgeons. To address your question, let me make note of that and see if we can get data for you when we come back.

DR. WEISS: Glenda.

MS. SUCH: Glenda Such here. Aside from thanking Dr. Weiss for bringing up the concern about what activities besides boxing and basketball or whatever a consumer might want to avoid doing, I believe during one of the discussions from the presenter we heard that nighttime activities that would be affected aside from having starbursts and halos during driving, one of the presenters actually had said the word newspaper print. I was wondering what other types of activities or type of events you've actually noticed being hindered by this lens with low illumination?

MR. McCARLEY: We haven't had any reports as the sponsor from a site or from a patient that they have been hindered in a nighttime activity from having glare or halos or starbursts or any other visual effect.

MS. SUCH: Not during driving either?

MR. McCARLEY: That's correct. None that inhibit them from doing that.

DR. WEISS: One question in terms of the induced astigmatism, Doyle. You had mentioned that this was most likely from the wound. I assume corneal topographies were done to just confirm that anyone with astigmatism or maybe in some patients that, indeed, it was wound induced?

DR. STULTING: Corneal topography was not part of the protocol.

DR. WEISS: Okay. So it was sort of more the assumption or it went along with the refraction where the astigmatism was and where the wound was placed?

DR. STULTING: Right. We have refractions before and after and vector analyses to look at the astigmatic change from one to another.

DR. WEISS: And the astigmatic change would be consistent with the placement of the wound or did anyone -- is this just an assumption or did anyone actually look at it?

DR. STULTING: We didn't look at it probably in a sufficiently organized way to really address that.

DR. WEISS: Okay.

DR. STULTING: I think it was sort of believed that the investigators were sufficiently familiar with wound placement for cataract surgery that they understood what it would do.

DR. WEISS: Thank you. I want to thank the sponsor and we are now going to go on to the FDA presentation.

DR. TOY: Good morning, panel members. I'm Jeff Toy, the team leader for this PMA P030028, phakic IOL for the correction myopia. The sponsor has already given an excellent introduction of the results and a description of the device so I only have two slides to add.

This first slide is just to acknowledge the PMA review team. They did a good job of expediting the review of this PMA. The team members are Don Calogero, Carol Clayton, Gerry Gray, Susan Gouge, Sue Jones, Bernard Lepri, T.C. Lu, Elizabeth Riegel, and Pam Reynolds.

Second slide is just the order of speakers for FDA presentation. Dr. Lepri will be first and giving summary of the clinical results and posing the question to the panel, and Dr. Gray will be second with the statistical analysis of the endothelial cell count. Thank you.

DR. WEISS: Thank you, Dr. Toy.

Dr. Lepri.

DR. LEPRI: Good morning, members of the panel, FDA colleagues and guests. In my presentation this morning I will just present to you some highlights that you will need for consideration for making your recommendations today.

This panel has specific goals to achieve today and those will be for us to assess, evaluate, and identify. We'll be assessing the risks and benefits and evaluating the effectiveness and safety outcomes presented by the sponsor and the PMA and their presentation here today.

Some of the risks that we've identified are operative and postoperative. Operative risks may include improper enclavation leading to surgical repositioning, wound leakage, infection, induced cataract and/or corneal damage due to surgical trauma.

Postoperatively one may see elevation of IOP inflammatory responses, the potential for pigmentary glaucoma as a result of iris irritation, critical losses of corneal endothelial cells and function, retinal detachment and dismemberment of the IOL itself with concomitant optical side effects such as glare and halos, etc.

Correction of high refractive errors without the optical limitations imposed by spectacles and the complications of long-term wear contact lenses is perhaps the major benefit for the patient, while reversibility and expanded options for treatment of high-refractive errors benefit both the practitioner and the patient.

I'm going to give you a capsule view of the effectiveness and safety outcomes that were presented here today. Under effectiveness some major highlights are UCVA, BSCVA, predictability of RSE, and the stability of the MRSE.

Uncorrected visual acuity of 20/20 or better was achieved by more than 30 percent of the overall treated subject population at one, two, and three years. UCVA of 20/40 or better were achieved by greater proportions ranging from 84 percent up to 87 percent over the three-year period reported in the study.

As one would expect, BSCVA shows that at least 79 percent have 20/20 or better and essentially 100 percent had BSCVA of 20/40 or better in the overall treated population. The ARTISAN showed a high degree of predictability in targeting refractive correction. At least 72 percent were within a half diopter of intended correction and 94 percent and higher were within 1 diopter.

At present refractor procedure stability is determined by evaluating the proportion of eyes that show variability in refraction no greater than 1 diopter between consecutive visits and refractions at least three months apart and mean differences of less than .5 diopter over a yearly interval.

The ARTISAN study population showed 95 to 98 percent were within 1 diopter of refractive change between consecutive refractions and mean differences in refraction ranged only from -.02 to -.05.

Safety issues where the BSCVA, which was already discussed, induced astigmatism, cells and flare, corneal edema, increased IOP or glaucoma, cataracts, and endothelial cell loss and corneal compromise.

Induced astigmatism of 2 diopters or more was reported in proportions ranging from two percent to 3.5 percent and the established target for refractive procedures has been set for less than 5 percent. The rates of inflammatory responses postoperatively were in the expected ranges that one would expect for this type of surgery.

While there were several reports of elevated IOP none persisted beyond 20 days post-op and were secondary to either postoperative steroid treatment or a few cases of incompletely aspirated viscoelastic. The cases that require short-term treatment all responded adequately.

While there were 49 lens opacities reported in the study only four were visually and clinically significant. The others were due to careful observations on the part of the investigators identifying normal age related chances in the crystalline lens. And of the visually significant cataracts three required extraction and the fourth one resulted in a loss of two lines of BSCVA but, to the best of my knowledge, was not worse than 20/40.

While there were no cases of actual corneal compromise reported during the investigation, endothelial cell loss changes were reported during both the short term in the domestic study and in the scant but long-term data from the European study. Dr. Gray will present the detailed analysis of these changes following my presentation.

I'm going to ask you to identify thresholds of critical inclusion criteria to minimize risks and perhaps the population it may benefit most. With considerations for the outcomes presented to you by the sponsor here today and in the PMA, the panel will be asked to make these recommendations regarding patient selection criteria, the risk benefit ratio of this device, and its associated surgical procedure and to establish criteria for product labeling if approved for marketing.

The use of phakic IOLs for the correction of refractive errors shows concern for the long-term effects upon the integrity of the corneal endothelium. The entry criterion established by this sponsor at the inception of this study was a minimum pre-op cell count of greater than or equal to 2,000 cells.

On the next slide and on slide 32 I need to make a correction. The mean pre-op starting is 2,754 and not 2,500 as on the copy of the slides that you have in front of you.

The sponsor's response to FDA's challenge of endothelial cell change data outcomes resulted in the sponsor's development of the charts you see presented here in this slide. Assuming a baseline cell count of 2,754 cells and assuming linear loss over time, the sponsor shows that after 30 years the cell count may drop to 1,272 cell per millimeter-squared. Of course, it is important for us to keep in mind the large margin of error viewed by spectral microscopy and the mathematical assumption of linearity and cell changes over time in these calculations.

The very nature of endothelial cell examination and change is affected by many variables. One variable identified in this study was anterior chamber depth. While the same size was low, it is particularly relevant to the ARTISAN lens its position in the anterior chamber and one can see from the six-month post-op period to three years for the seven eyes having anterior chambers ranging from 3.0 to 3.2 mm that there was an estimated cell loss of 8.99 percent.

The ARTISAN also offers two models whose optic sizes vary. They are 5 mm and 6 mm and relate to the patient's pupil sizes. The relevance of these optic sizes is related to performance in low-light environments and the potential for symptoms and complaints of glare and halos that may impact functioning such as in nighttime driving. The sponsor presented the outcomes of patient satisfaction by questionnaire responses for our consideration.

The implied refractive benefits of the ARTISAN have already been discussed here today and are directly related to the targeted refractive range. You will recall that only a small percentage of eyes were treated below -8 diopters of myopia.

I am not going to present the questions now. We will present the actual questions to you following Dr. Gray's presentation of the endothelial cell data. Thank you.

DR. WEISS: Thank you, Dr. Lepri.

DR. GRAY: Good morning. My name is Gerry Gray and I'm going to discuss the results from the endothelial cell counts in this study. I'm the team leader for the Cardiovascular and Ophthalmic Statistics Team. This submission was mainly reviewed by a member of our team, T. C. Lu.

Just a synopsis of what we're going to be talking about here. The purpose of the endothelial cell count is to investigate the effects of the device on the endothelial cells through time. We have endothelial cell counts and measurements from specular microscope photographs. There are multiple images for eye after all 2,000. We have counts at baseline six months one, two, and three years.

As you've already heard in some detail from the sponsor, there was a very large variability in the initial set of data and so images were reread as possible and the net result is we have 353 available eyes from reliable machines that were recounted in one reading center. That was a total of 1,144 actual observations eyes by visit. As a statistician I need to point out that we don't have any control group here so it's very difficult to evaluate the results without an actual control.

So there is no control and the question is what do we compare these results to? We want reasonable assurance that the endothelial cell density is preserved. The normal loss due to aging is apparently around 0.6 percent per year. The point for concern appears to be around 1,000 to 1,200 cells per millimeter-squared.

There are several sources of guidance or preliminary guidance and they are all written in terms of trying to place an upper confidence limit on this rate of loss. The FDA draft guidance and the discussion from this panel several years ago set an annual rate from three months to three years, an upper 90 percent confidence limit of 1.5 percent.

The ISO and ANSI documents are not actually, I don't think, written in terms of standards for acceptable rate of loss but those both suggest you calculate a sample size for this kind of study using a 2.0 upper 90 percent confidence interval.

Here is a visual representation of the data that we do have from a recount study. Each vertical bar is one of the visit, base line six months, one, two, and three years. The green indicates that we have actually a count in that time and the white indicates we don't.

Individual eyes can be read horizontally across here. Here on the bottom are the 57 eyes that were measured at all time points. There were 126 extra eyes that had a baseline measurement and by the end there is actually 50 of them left here so there's 107 eyes that have both base line and three-year measurements.

Then there's a fairly large portion of eyes, 170 right here that have no baseline measurement. Then you can see these numbers indicate the number of people that started in at those various points in time.

A couple of comments on this graph. This is not the normal pattern. We are used to dealing with

-- this is not the normal pattern of missing we see where initially everyone has a baseline and people drop out through time.

This is somewhat unusual because we have this very large group that actually doesn't have measurements at the beginning. That was, I'm pretty sure, due to the fact that the study was sort of given a lot more importance part way through. Initially we don't have baseline measurements for these people.

Another comment that I want to make that came from the discussion, the earlier questions, the question of is there a bias problem because perhaps some of the measurements are thrown out because they were low. The question is how is that going to change the rate of loss through time if there's a bias?

What I want to point out is that if there's a bias, there's more people missing here at the beginning than there are at the end so I'm not sure if there's a bias how it would affect any kind of results we have here today. I think that's an unanswerable question.

This is a plot of the actual data that we do have, the 1,140 observations from 353 eyes and the blue line just connects the means at each time point. The red line across the bottom, just for your reference, is 1,200 cells per millimeter-squared.

Now, what we're interested in is the steady state, if you want to call it that, the long-term loss that we can expect to see. That estimate depends on a lot of things. It depends on the model that we use, whether we account for an initial operative loss or not so the function of formal use, whether we use the baseline in the end or some form of regression, the cohort we use, the details of the statistics.

As an aside, it's not entirely clear to me that natural loss for untreated patients is actually steady state either. That further complicates any kind of extrapolation you want to make.

All that aside, there's really not that much variability in the estimates of long-term loss from these data. The sponsor presented an annual loss of 1.7 percent based on 183 eyes but had a baseline count. That calculation is based on a regression that includes the baseline. A 90 percent confidence interval for that is 1.3 to 2.1 percent.

An alternative that I think might be slightly better uses all the data that we do have and tried to account for the missing using something called multiple imputation. That actually gives a fairly similar result, 1.8 percent annual loss, 90 percent confidence interval 1.3, 82.2 percent. Both of these estimates account for correlation within patient in a reasonable way.

Here are the results from the best, the 1.8 percent loss per year. If you actually pull out the other one on top of this, the lines are virtually superimposed. It looks almost the same, 1,200 cells per millimeter still there as a reference.

Now, of course, this is what we have so far for three years and what you really are concerned about is what happens in 10, 20, 30, or 40 years so we want to do some extrapolation if we can. Before we do that, it's my duty to remind you that we are trying to -- it's always a questionable exercise to extrapolate and we are trying to extrapolate 10 times the range of the data that we do have.

All that being said, though, probably some type of -- you have to make some extrapolation to make a judgment, either formally or informally. If we do it formally, it's very dependent on the model we use and the assumptions we want to make, is it linear, exponential, whatever kind of decay.

The problem is with only three years of data we can't really distinguish between these models. There's no way of telling what happens if things change in 10 years. Because of that I think you also should really consider if it's necessary to obtain good long-term data and how you might want to go about that.

One more thing. We do have some additional long-term information that has been referenced previously. The sponsor has provided additional four-year data on 27 patients who showed a 1.63 percent loss between three and four years. Then there is some additional long-term information from a 19-patient European cohort.

Basically the same follow-up is in this study but there is an additional point t 10 years. For those patients their mean counts went from 2,666 to 2,180 at 10 years. That's an 18.1 percent decrease over the 10-year period. Six percent of that was in the first six months.

That translates into annual rates that you see down here at the bottom, 1.2 overall. The rate between six months and three years was actually fairly high, 2.9 percent, and the rate between three years and 10 years is actually fairly low, 0.7 percent. You can make what you will of that.

After we got all the caveats and other data aside, here is a picture of the linear extrapolation that you would produce using the 1.8 percent loss per year. On the graph are also confidence limits, the dash lines of the confidence limits on the regression and the dotted lines are the confidence limits for predicting an individual.

You can see there is a fair amount of variability and what really matters is right from here there is a fair amount of variability in this direction. In other words, the variability and the time the person might take to reach 1,200 cells per millimeter-squared. All this, of course, still assumes that whatever happened in the first three years is going to continue linearly for the next 37.

Using a linear model we can actually -- and using the rates of loss that we get based on the estimates we produce we can produce a table that shows the years until predicted 1,200 cells per millimeter-squared. You can see of you start out at 2,000 cells then after 12 to 17 years, depending on how cautious you want to be. you are going to be at around 1,200. If you start out with 3,200 cells, then you have maybe 30 or 40 years until you reach 1,200. Again, this all should be taken with a grain of salt because it's an extrapolation and there is a fair amount of error.

Maybe a little more important than the average cell loss through time is a question of how are the individual patients faring here. In other words, what proportion of the patients are going to show a cell loss that's greater than some critical amount. Another way to ask that is what proportion of patients are going to have cell densities less than 1,000 to 1,200 in 10, 20, or 30 years.

Again, it's hard to answer with much confidence because now we're not just extrapolating the mean. We are trying to extrapolate the percentiles. We want to know what's the lower 10 percent of the patients and where are they going to be in 10 years.

The best I can think of with the data we have is to take all the patients that we actually have. We can actually fit a regression. We have more than two observations on them, more than two follow-up visits. We can fit a model that actually gives each of them the possibility of having their individual rate of loss.

The model actually is called random effects regression. What it does is assumes that the losses come from some normal distribution so the rates of loss are coming from a common distribution. That's what you see here. These are the results. The dark lines indicate the 1.5 and 2.0 percent losses. You can see that most of them are below 1.5.

So also using that same histogram we can save the percentage of patients with annual losses worse than a particular amount what can we expect. Using these data and this model you can say that probably 5 percent of the patients are going to have losses of 2.2 percent or more, 99 percent of 1.5 percent or more.

Again, I need to give some comments on these estimates because they are fairly highly dependent on the model used to arrive at the individual patient estimates. The model, on the one hand, reduces the variability because it shrinks the estimates. The estimates for each patient are moved toward the overall mean so that reduces the variability and that would tend to make this number a little bit smaller.

On the other hand, the annual loss in this model where I didn't do the imputation was a little bit higher so that would tend to counteract that to some degree. This is the most I can give you right now.

Just to summarize, if I can, in one slide, the estimated annual loss is apparently about 1.8 percent per year with a 90 percent confidence interval 1.3 to 2.2. For individual patients maybe a third of them have annual rates of loss more than two and five percent have rates of loss more than 2.2. Again, it is necessary to do some form of long-term extrapolation but you need to try to interpret that with whatever amount of caution you want to put into it. Thank you.

DR. LEPRI: Okay. I'm going to present question 1 to you and then there are several slides of background data that you need for consideration. You have the copies in front of you. I'm able to put all of those charts into a slide form so you may want to refer back and forth to them.

The first question is:

1. Do the endothelial cell data presented above by overall analysis, stratified by anterior chamber depth and the extrapolations over time provide reasonable assurance of safety of the ARTISAN myopia lens?

Here is the data that was presented and the hardcopy questions that you have in front of you. The first slide shows the estimated changes in cell loss at six months, one year, two years, and three years. The standard deviations, errors, and confidence limits.

The next piece of information that you are to use is the percent change from baseline. It shows also for the intervals of six months through three years. The percent change by period, the difference between six months to one year, one year and two years.

In this slide it shows that in a paired analysis the percent change calculated between baseline and three years post-op was -4.76 percent with a standard deviation of 7.8 percent. When analyzed by interval one can see that losses appear to be higher between the second and third postoperative years.

The sponsor did show that when they eliminated the one site, that all had the specular microscopy done with the same device, when they had changed employees midstream during the study when they removed that data out, that dropped from -2.37 percent to minus 1.68 percent.

DR. WEISS: I would just request whoever has the cell phone if they could silence it forever. Thank you.

DR. LEPRI: The next slide shows the endothelial cell count change over time from baseline stratified by anterior chamber depth for the 3.0 to 3.2 mm anterior chamber depth. You can see the changes over time. Even though the ends are small, there is no statistical significance to this but we want it for consideration for potential trend.

The next slide is endothelial cell count changes from six months to three years stratified by all of the anterior chamber depths in the study. The last slide is the subjects with three and four-year follow-up having that mean ECC at pre-op of 2754 with an end of 27 to show what their changes were from three to four years.

2. Do the other data presented in the PMA outside other endothelial cell data provide reasonable assurance of safety? Those are to be considered as two separate issues.

This is the background for Question 3. The proposed statement of indication reads: "The reduction or elimination of myopia in adults with myopia ranging from greater than -5 to less than -20D with less than 2D of astigmatism at the spectacle plane; Patients with documented stability of refraction for the prior six months, as demonstrated by a spherical equivalent change of less than or equal to 0.50D."

3(a). Does the panel recommend any modifications to the proposed statement of indications with respect to:

a). minimum anterior chamber depth;

b). maximum pupil size (the 2 models of the ARTISAN are intended for patients with pupil sizes up to 5.0 mm and up to 6.0 mm; and

c). minimum preoperative endothelial cell density? The outcomes of ECC changes reported in the background data for question No. 1 above should be referenced if the panel wishes to recommend an acceptable minimum endothelial cell density to quality a patient.

4. Do the panel members have any additional labeling recommendations?

DR. WEISS: Thank you very much. We are actually doing fairly well on time so what I would ask is if the -- I hear chuckles. I guess usually we haven't been in the recent past. What we're going to do is if the FDA could perhaps entertain some questions before lunch and I'm going to ask if anyone from the panel has any questions.

DR. BRADLEY: I have a quick question for Dr. Gray.

DR. WEISS: By the way, I wanted to thank Dr. Gray for that wonderful first slide showing where people fell out in terms of participating and not participating in specular microscopy because that really just clarified things amazingly.

DR. BRADLEY: Dr. Gray has presented a similar presentation some time ago, if I recall, to this group. In both presentations you have admonished us to be very aware of the shortcomings of extrapolation. In spite of that, we go ahead and extrapolate primarily because most of us are not very sophisticated. I think you always give us a linear model which we can sort of understand because we can all draw a straight line with a rule.

But in the end, from my perspective as a scientist not involved in this field, I just find myself incredibly uncomfortable with this extrapolation and I wondered do you know of any data from some other product, other condition that indicates that the pattern of cell loss seen in the first three years is, in fact, continued on in a linear way over five, 10, 15, or whatever years? I don't know this field at all and maybe you could help.

DR. GRAY: Well, first of all, you might have noticed that I said in this presentation that some amount of extrapolation is necessary to make a decision. Even though it's my job to warn you about it, you still have to do it.

In terms of further data that might corroborate any kind of model, all that I know about is what we presented in the 19-patient European cohort. I actually, if you really want to see it, I have a plot somewhere. If you plot those 19 patients superimposed on the extrapolation, they basically cover the whole range of error for prediction of an individual. They are right there. There's only 19 of them and when you look at that they have a fairly large amount of variability so it doesn't really help us to decide sort of a relatively subtle difference between something like a linear loss or an explanational loss or something like that.

DR. BRADLEY: Thank you. I'll open the question up to anybody else in the room who is knowledgeable in the issue of endothelial cell count data. Are there any data for some other product, some other disease that we have long-term data on?

DR. WEISS: Dr. Grimmett.

DR. GRIMMETT: Dr. Michael Grimmett. In my review of endothelial data for this panel perhaps a year ago, the only other data that I could find would be Bill Bourne's data. His data had several limitations in that the patients that had the cataract surgery had a wide variety of the types of procedure whether it be extracap or intracap.

Specular microscopy images were not standardized. I don't believe that the Konan machine was around at that time. I didn't go back through the data to look at it year by year to answer your question did the first three years actually predict what happened 10 years later. That's the question you're asking. But his data was such small numbers and such a wide variety of procedures that I'm not sure that would actually even looking at his data would actually answer it. From my review I'm not aware of another product where we have the answer to that question.

DR. GRAY: Here is the trial I was referring to where the red dots have the 10-year European data. You can see they neither confirm nor deny anything about -- their variability is fairly large here in these 19 patients and so they don't really tell me that the model is terribly wrong but they don't help me distinguish between fairly subtle differences.

DR. WEISS: Dr. Huang and then Dr. McMahon.

DR. HUANG: I know we spend a lot of time on endothelial cell counts from the FDA as well as the panel reviewers as well as the sponsor. I would like to look at this problem with a little bit slightly different angle. Truthfully that the cornea function is not really predicated on the absolute number of the endothelial cells.

It's really their functions. So are we looking at the cells as indicative of function or should we just look at the cornea thickness as a function to see if the cornea retains integrity because clinically we have seen many patients with endothelial dystrophy with reduced cell count but over the years they don't have any cornea decompensation. Even though the cell number continues to decrease, that doesn't mean the cornea is decompensating. That is my concern about all these number calculations. I understand that we need to have safety guidelines but, on the other hand, that's the only safety guideline that we need to be concerned about cornea integrity. Thank you.

DR. WEISS: I think the difficulty will be that the cell count is going to be much more sensitive, perhaps not totally significant, than the corneal thickness because as we all know as corneal surgeons, the thickness or the cornea will decompensate at a much lower cell rate.

If you are a 20-year-old patient and let's say you're losing your cells at 3 percent per year, and it's linear and continual, then we would obviously have concerns at some point. You may get into the risk of having decreased corneal function. These are all very difficult questions because I think what we're being told by FDA and by sponsor we have a 1.7 to 1.8 percent corneal endothelial cell rate loss in the first three years.

It doesn't stabilize. What we all know is the only time this will become significant is many, many years down the line past when hopefully all of us will be retired at that point and not meeting at this panel meeting but we need to project into the future with data that we don't have.

Dr. McMahon.

DR. McMAHON: This goes back to Dr. Gray. This might be extraspeculative but in that European data is it possible to use a nonlinear model? The issue here is there a decrease in the rate of change at the end that would show some flattening? I mean, the plots that you show demonstrate that these individuals if this is real are doomed if they live long enough.

DR. GRAY: It's possible to fit a nonlinear model but it's impossible with the data we have to distinguish between a linear or a nonlinear model. We can do those fits if you want to extrapolate in some other way with some other model, you can either make it curve one way or the other and look either better or worse. I have no basis based on the data we have to pick one of those models over the other.

What I present here is just the straight line middle-of-the-road linear extrapolation. If you have some reason to choose otherwise, we can entertain another model. It's difficult. It's impossible with the data we have, I think, to distinguish between those.

DR. WEISS: Any other questions from panel?

DR. BRADLEY: Sorry, Dr. Gray. You stepped down. I'm still not clear on what you've shown us here. The red dots --

DR. GRAY: Is this on?

DR. BRADLEY: Let me get my question out and you can answer it. For example, these are 10-year follow-up. Presumably these people at this time are 10 years older and one wonders what the age match norms might be for this group. That looks to me like most definitely the means must be lower in this sample that you put up there, the 10-year follow-up.

I wonder how different are they to age-matched controls, age-matched norms, for that group of people whatever age they were. I'm trying to get a sense does this group really have lower than normal looking endothelial counts. That wasn't a very clear question. Sorry about that.

DR. GRAY: Well, first of all, let me make it clear that I did not do -- these red points were not included in making this fit at all because I didn't -- I don't have enough information to have any idea whether we can pull together the data and use them in the same model or not. This plot was only made just in case we wanted to see how it looked instead of looking at the figures that I presented in slide No. 10.

Again, all I had, I personally got these data last week so I didn't have a lot of time to fiddle with them. All I had was the -- I don't have the co-variates. I don't know their ages. I don't know anything about them. I don't know the pupil diameter, none of that stuff. All I know is -- all I got was the counts at baseline and the various follow-ups.

In the 10-year European, the slide that had that was just to indicate it. This is all we really know about long-term. This is the best we have in terms of long-term follow-up. This plot is just another way to look at that to see if there was some obvious red flag that any kind of extrapolation was off the mark. Really what the plot tells you is that there's not much information here.

DR. WEISS: Dr. Mathers.

DR. MATHERS: Bill Mathers. What you're saying is that those red dots are actually extraneous to this graph. They happen to fall right down the middle where the extrapolation is which would mean that the extrapolation seems to be consistent with the 10-year data of the European but, of course, you can't really say that.

DR. GRAY: I would say it's not inconsistent.

DR. MATHERS: It's not inconsistent.

DR. GRAY: I'm a statistician. But also there are some patterns in the European data that are different than the data we see here. For example, 353-eye cohort that we looked at there was virtually no change between baseline and the six-month follow-up which is counter to anything I have been led to expect.

Whereas for this European cohort there was a six percent loss between baseline and six months. So the patterns even though it comes out the same in the end at the 10-year point. The patterns up here at the beginning are somewhat different. Who knows if it's just due to the few number of patients or that they are really different patients. The population is somehow different demographically. I don't have that information.

DR. MATHERS: But to the subjective eye it looks like those red dots were used to calculate it because they look smack on.

DR. GRAY: They do but you will also remember that I mentioned I think it's three or four of them above the 90 percent line and four or five of them are below. They actually have a fairly large amount of variability compared to the line that we do have.

I don't know how they got these counts. I don't know how the counts were standardized or anything but the amount of variability is actually fairly large here compared to what we had seen before in the current data set.

DR. WEISS: Is there a zero timeline for the European data? We have it on the 10-year.

DR. GRAY: If you look at slide 10 at baseline, there was 2,666 which was 100 cells lower than the mean and about 100 cells lower than the 2,760 in the current cohort so they started out slightly lower.

DR. WEISS: So just following up with what Dr. Mathers is asking, if that was plotted out there, would that fall quite similarly with the black line?

DR. GRAY: If you look at --

DR. WEISS: That would sort of correlate with what Bill is asking, that if it looks similar at zero and it looks similar at 10, then maybe it actually --

DR. GRAY: The change --

DR. WEISS: Maybe it's not inconsistent with being similar.

DR. GRAY: Actually, the change for the -- I didn't want to make too much of -- we only have 19 patients and I don't know much about them but, having said that, for that cohort the average loss between six months and 10 years, the annual rate is 1.2 percent. It's actually lower than what we saw in the PMA cohort.

They had a very large drop at the beginning and then they leveled out somewhat. If you look at slide No. 10 it has a whole bunch of different ways of looking at the data to try to help you make some sense of that.

DR. WEISS: In the European data they only had 19 patients and there was a large amount of variability so all of these are deficits of over analyzing this data. Having said that, they have a 1.2 percent cell loss rate. Okay, good. From six months to 10 years.

DR. GRAY: They had a fairly high rate of loss between six months and three years, 2.9 percent. It was high. And then between the two time points, three years and 10 years, it dropped off to 0.7 percent. If you are optimistic you say the long-term rate is close to normal. If you are pessimistic you say the initial rate in the first three years was quite high and I don't really believe -- there's not enough data here to really tell what is going on so it's a judgment at this point with those 19 patients in my opinion.

DR. WEISS: Dr. Macsai.

DR. MACSAI: Dr. Gray, can you address something about this slide? I thought enrollment criteria was 2,000 cells or above. On the slide at the zero there's a whole bunch of little points. Maybe it's my refraction. I can't see how many little points but they are below 2,000.

DR. WEISS: You need to get an ARTISAN.

DR. MACSAI: My contrast, I think.

DR. WEISS: Sorry. Getting close to lunch.

DR. MACSAI: It seems like there are little dots on your graph below 2,000 at baseline.

DR. GRAY: There are.

DR. MACSAI: How is that possible?

DR. GRAY: Well, it looks to me like there's four or five dots below baseline at 2,000. You will recall that these are the recount data. These are not the initial counts so it could have been that when the patient was enrolled whoever did the endothelial cell count deciding it counted them one way, and you will remember there is a fairly large variability in the counting process so it's not surprising that a few of them actually came out lower when you recounted them. That's why the new suggestion is three photographs per person and standardization of the counting procedure to try to minimize that kind of variability.

DR. MACSAI: So we're not even 100 percent certain that our baseline counts, because these are based on one picture where all of those below 70 were kind of thrown out and we don't even know if that amount was thrown out was randomly distributed or skewed in some way. We don't even know if our baseline is right is what you're saying in a statistical manner. I mean, where you don't want to be committal but that's what it sounds like.

DR. GRAY: What I'm saying is the sponsor had a slide that talked about the about of variability in the measurement of the endothelial cell density. There actually is inherent in this whole process a fair amount of variability. We take photographs of some location in your eye that can vary. Some of the photographs turn out good or bad for whatever reason and then we have people trying to count and to obtain a density, a cell density.

Just that whole process has a fair amount of variability in it. When you say sure, we're not positive of any of these counts. They have some measuring error. The recount data have less variability than the original study.

DR. MACSAI: Based on what do you say that? I mean, there's no standardization. It sounds like there's no check and balance done before it started.

DR. ROSENTHAL: Can I just explain something?

DR. MACSAI: Yeah. I'm really confused.

DR. ROSENTHAL: Their initial endothelial cell counts were done with large -- were not done in the standardized way. They were all over the board when it came to the variability. The Agency asked them to go back and to try out of this large number of eyes to get those that were taken standardly, were counted standardly, and were evaluated standardly. It's the best, frankly, I think we can do particularly when a new modality to look at the endothelial cell counts came up in the middle of their study.

DR. WEISS: Dr. Schein. Sorry.

DR. ROSENTHAL: They were all using different methods of doing it.

DR. WEISS: Dr. Macsai wanted to follow up and then Dr. Schein and then Dr. Grimmett.

DR. MACSAI: I feel an obligation here to make a follow-up statement, Dr. Rosenthal.

DR. ROSENTHAL: Sure.

DR. MACSAI: I believe that the Konan specular microscope was available in 1997. Whether or not someone chose to utilize it it existed. Let's not preclude that it came about in 1999. That's point No. 1.

Point No. 2, from our history as ophthalmologists knowing the complications of anterior chamber intraocular lenses in patients, the Lysky, the ORC, when we designed these studies using an ACIOL I think it behooves the sponsor and the Agency to address these critical issues at the beginning before we move forward with implantation in patients because now we're looking at maybes.

DR. WEISS: Dr. Rosenthal.

DR. ROSENTHAL: I have to stick up for the Agency a little bit. I think in 1997 there was not as great a science of endothelial cell count as there is in the past three or four years. Certainly working on it in the standards group it was a very contentious issue and it took a long time to come to some conclusion how best to do it.

I don't know if Donna wants to comment on that. When a company puts together a protocol for an IDE, we have to use what is currently considered the best science. Frankly, the science of endothelial cell counts in 1997 did not have a quality standard.

DR. WEISS: That will be the last word on that subject. I would like to go back to questioning. We just have a few minutes right now. Dr. Schein, if you have anything that you -- a question as opposed to any comments.

DR. SCHEIN: You've taken a comment right out of my mouth but I have one last question for Dr. Gray. Putting the cornea aside for the moment, I'm interested to know if you did any time dependent analyses of other complications, development of lens opacities, need for cataract surgery, intraocular lens or lens exchange, retinal detachment, etc., etc., both within the time frames of the data that you have and an extrapolation into the future.

DR. GRAY: The brief answer to that is no, I didn't do any of those analyses.

DR. SCHEIN: I would suggest they might be useful if for nothing else than patient education to describe whether if you survive the first month or year or 18 months, that the complication rate goes down dramatically or the converse obviously equally important.

DR. WEISS: Fifty seconds.

DR. BRADLEY: Dr. Bradley. Again, Dr. Gray, question from your analysis. Did you notice whether the cell-loss rates correlated with the initial cell count.

DR. GRAY: As far as I could tell they did not. There was no significant indication that the rate of loss was a function of the baseline count.

DR. BRADLEY: So would the appropriate interpretation of that result be those with the low initial cell counts are at the greatest risk?

DR. GRAY: Yeah, I would say that's a fair interpretation of that.

DR. WEISS: Depending on how long --

DR. GRAY: As far as I recall, there was not -- it's difficult to work with the data when a lot is missing like this but I couldn't find any association between the baseline count and the rate. As far as I can tell the best thing to do is just assume that it isn't a function of the rate and if you are low to begin with, you're at a higher risk.

DR. WEISS: Thank you very much. 12:30. We'll break for lunch for one hour.

(Whereupon, at 12:29 p.m. off the record until 1:36 p.m.)

A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N

1:36 p.m.

DR. WEISS: Okay. So what I would like to do before we go to -- we are going to be starting -- before we start committee deliberations, I had one question for the FDA which was on the basis of their presentation if they any recommendations as far as a time point after lens implantation, which would make it much easier to extrapolate, the endothelial cell count some years down the line as opposed to having to wait 20 years to find out what the answer would be in 20 years. I don't know who would be able to answer that one for us.

DR. GRAY: I can give you my opinion on that.

DR. WEISS: That's the one we want.

DR. GRAY: What you're trying to do is extrapolate 10 times the range of the data that you have. That makes any kind of distinguishing between

-- several models could probably fit equally well within the relatively short amount of time you have, three, even if we have four years, and still be fairly divergent after 30 or 40 more years.

It's going to be very difficult in terms of extrapolating out 40 years and know anything until we do get to the 10 or 20-year point. That's obviously somewhat impractical in terms of making a decision about approval.

Every year helps. Every year that you have further on that has no obvious increase and perhaps a decrease the better off you are. You are never going to be able to prior to approval have enough data to definitively say that it's one particular kind of functional form as far out as you want to go.endothelial

DR. WEISS: So from what I understand you to say that if we have four-year data or five-year data, that would not make the answer anymore clear than having three-year data.

DR. GRAY: In terms of the extrapolation I don't know that it would make that much difference in terms of distinguishing between a straight line and a curve, something like that.

DR. WEISS: Okay. Thank you.

There were a few questions that we had asked sponsor to look up. I'm told they have the answers to some of these. If they could come forward. There was a question I had about pupil size and explantation and a question that Dr. Casey had and Dr. Smith had.

DR. STULTING: Thank you, Dr. Weiss. We worked on this during the lunch break and I'll share with you the data that I have. There may be some more available later in the day. One of the questions that I may note of was the issue of mesopic pupil size and lens optic size. The sponsor did a multi-variate analysis looking at the presence of visual symptoms at night looking for correlations.

One of the correlations that they sought was mesopic pupil size greater than the lens optic size. In the cohort there were 56 first eyes enrolled and 31 who answered the questionnaire who fit this criterion. There was no correlation found in that analysis. I don't have power calculations available. That is something we can get for you later.

The second question that I made note of was some concern about the possibility of bias in the selection of recount patients. I want to spend just a minute going over the protocol that was used to select those eyes.

The selection of sites for the recount was based only on the availability of instrumentation. It is possible that there is some unrecognized bias that people who are particularly good surgeons happen to have particularly good specular microscopes or something like that that we can't definitively and absolutely rule out, but there was no intent for that.

All available readable images regardless of endothelial cell morphology were included. In fact, this was a masked selection. The images were read -- were obtained and read at a central center not knowing who they belonged to, whether they were preoperative or postoperative, etc.

Once they were read, then a minimum of two readable images at different time points was required in order for an individual to be a member of the recount study. The other question that was related that was asked was how many images with only a few cells were eliminated? The answer to that is there were 12 poor quality images eliminated because there were less than 70 analyzable cells in those images.

Those were the exclusions among 1,156 images that were analyzed leaving a total of 1,144 images which formed the data set that the recounts were derived from. We believe that the elimination of these few images probably didn't have anything to do with the results.

The third question was endothelial cell counts for Group E. Remember Group E was the group with replacement intraocular lenses, previous corneal transplants, custom made lenses that were fabricated with powers outside of the usual range, best corrected acuities less than 20/40.

Nine of these were included in the recount analysis. Three of them had replacement intraocular lenses. Two of them had custom lenses. Four of them had best corrected acuity of less than 20/40. There were 23 observations in this group so it was a relatively small group and in these there was an average loss of 2.67 percent per year. Recognize that one-third of these were people who had had an extra surgical procedure to remove the intraocular lens.

A question was asked about endothelial cell count reliability. I answered it by saying that the protocol did not have any internal controls for reproducability. However, I would like to share with you some data about endothelial cell count reliability since the question was asked.

Once the images had been obtained, screened and read at a single trained central center, those images -- 50 of those images were randomly selected and sent to another reading center. This is a center that was outside of the investigational sites and a center that most of you would probably recognize that normally does endothelial cell counts.

So these same images were read by the second center. This then is a test of reproducability of reading alone because they were exactly the same images. The differences un the mean cell counts in this exercise was 0.8 percent, not significantly different from zero. But the standard deviation was relatively large, 24 percent, ranging from -47.2 to +48.8 percent and 28 percent of these readings showed a more than 10 percent loss or gain. So this speaks to the ability to read these images. I speaks to the reliability of the methodology for endothelial cell counts.

Remember that these cells -- we are only counting 80 to 100 cells in most of these eyes, mean 109 even with selected images. If you are off by two or three cells, it makes a big difference in the calculated endothelial cell density.

With regard to the labeling, I would just like to make a suggestion and that is that we produce a graph something like this showing a calculated endothelial cell loss over time and relating the endothelial cell density to the age with endothelial cell density on the vertical axis and age on the horizontal axis using our best data available with the best projects of time so that I as a consumer, as an ethical physician, can have this information knowing that it would be best to implant or not implant depending upon these parameters. Thank you.

DR. WEISS: Thank you. We're going to go on with the primary panel reviews. Dr. Mathers.

DR. MATHERS: Thank you, Dr. Weiss. Bill Mathers. I will relate to you my primary review. The application concerns a lens that is designed to correct myopia, moderate to high degree, five to 20 diopters by means of a lens device that is inserted into the anterior chamber and clipped to the anterior surface of the iris which maintains it's fixation and it's centration.

The highly myopic population has significant problems with spectacle correction. Contact lens are usually the preferred method of correction for this group if they are tolerated. Subject with dry eyes, surface disease, and other difficulties that preclude contact lens wear have few options.

We are given the question for the panel discussion, "Do the endothelial cell data presented in the overall analysis stratified by anterior chamber depth and extrapolated over time provide reasonable assurance of safety for the ARTISAN myopic lens?"

There are several safety considerations that need to be addressed. The primary and overriding issue, however, is, I believe, the question of endothelial cell loss over time and the change in endothelial cell density resulting from the insertion and retention of the lens.

Data supplied by the applicant is presented in two forms, for the whole group and for smaller subgroups stratified by anterior chamber depth. For the whole group the endothelial loss rate for three years, the duration of the study was 4.75 percent and this is a loss rate of 1.58 percent per year with an N of 111. I realize my numbers are not exactly the same as some others that we've heard but I believe actually they have come up pretty close.

This contrast with the loss rate in the normal population of .6 percent and a loss rate of 2.5 percent for 10 years following cataract surgery. This cumulative endothelial loss is highly relevant to the younger population for which this lens is primarily intended. The table below indicates the resulting endothelial cell counts that could be expected if the loss continues at this rate for 10, 20, 30, or 40 years. I realize you may not have that in front of you but I'm going to go over the numbers.

Starting with 2,754 cells per square millimeter the mean endothelial cell density may seem reasonable but half the group will have an ECD less than this. The applicant has requested permission to use the device in 21-year-old subjects with an ECD down to 2,000. The main corneal clarity usually requires -- to maintain corneal clarity usually requires an ECD of 800. These are rough figures but they are probably correct.

For a reasonable margin of safety an ECD of 1,200 would be a better cutoff and even this is fairly low. Starting from the mean ECD and the lowest cell loss rate the average subject would be at risk after 40 years. Subjects with an initial ECD of 2,400, usually considered to be quite good cell count, would reach the point of risk at about 30 years.

If the subject had an ECD at the low end of 2,000, the 1,200 end point would be obtained in 23 years and the 800 ECD would be reached before 40 years. A cell count of 1,200 does not guarantee imminent corneal failure but there is definitely an increased risk at this point. One needs to consider that these patients at that time are going to be facing cataract surgery which always has some consequence for the endothelium.

The data actually shows that the estimated loss rate from six months to three years, which is a total of 30 months, if this data is correct, then the loss rate is more like 1.9 percent per year. The resulting calculations shown above indicate that even starting with relatively high ECD of 2,754 the final ECD reaches 1,200 prior to 30 years. By 40 years the endothelial cell count is so low as to guarantee failure.

These calculations are based on a mean loss rate. Also given our 95 percent confidence internal which have a high-end loss rate of 6.1 percent for three years, or 2.03 percent per year. At this rate the 1,200 ECD is reached in about 26 years starting from the high end of 2,754.

Five percent may fall beyond this range and an ECD of 1,200 reached sooner than that. Starting from an ECD of 2,000, which they are requesting, the 800 level is reached before 30 years. I want to point out here from today's discussion that Dr. Gray's assessment at 38 percent of the population could be expected to have a loss rate of two percent which is a failure rate, or 1,200 rate at only 25 years.

The highest loss rate was found in a group with an anterior chamber depth of 3 to 3.2. For this group the loss over three years, or maybe 30 months, I'm not sure, was 9.16 percent or 3 percent per year. This is a loss rate that is approximately double the group as a whole. Thus, the time to reach 1,200 or 800 is half the original calculation.

From an ECD of 2,000 less than 20 years would be required to reach 800. These calculations assume that the endothelial loss is close to the mean. Unfortunately, this is not likely to be the case since the standard deviation reported in the revised application is nearly twice the mean number. This indicates that some subjects will likely experience a substantially more rapid decline in their endothelial cell density than these calculations show.

This device is currently marketed in over 40 countries and the report states that the device has not been removed from any of these for any safety concerns. This is not surprising because the time to achieve is sufficiently low ECD that would create corneal edema is still always over 15 years. Our 10-year data given to us before and also reanalyzed today I would think does not contraindicate or contradict this conclusion.

The endothelial cell losses are mostly less than those that have been reported for cataract surgery. A comparison with cataract surgery is relevant since clear lens extraction is one alternative that some practitioners use to correct extreme myopia.

For both operations there is a small incision into the anterior chamber and the device is implanted. Surgical trauma and postoperative inflammation could be expected to be of a similar range.

Cataract surgery is extremely common and the risks are generally considered to be low and reasonable. Why is this different here? The age of the cataract surgery population is higher and, thus, the postoperative duration is much longer for the ARTISAN myopic lens.

In addition, preoperative vision loss is greater for the cataract group and the relative risk of surgery can be correspondingly greater. Finally, there are alternatives to phakic lens implants, whereas a cataract patient requires the replacement of the lens to restore vision in this new alternative.

Other safety concerns of shorter duration, less than 5 percent of subjects lost two lines of best corrected vision and 100 percent at three years had a best corrected of 20/40 or better within 228. This is in the range of cataract surgery where severe vision loss can be expected in the 1,000 to 2,000 or less range.

One subject was developing PSE cataract and we heard some other issues about cataract formation today that I'm not quoting. There was one case of a macular hole. Over time the incidence of cataract may be higher because of subclinical inflammation from the lens.

But the rate of cataract development in this group is already higher than average and it will very difficult to make this attribution accurately. Postoperative inflammation in the form of cell and flare is persistent in 1.3 percent of subjects at six months.

This chronic inflammation may contribute to the cataract formation later. Corneal edema was surprisingly prevalent at 20 percent on day one and this dropped 2.2 percent in two weeks but I think this level is acceptable.

Regarding accuracy issues, the accuracy of the implant appears to be excellent considering the very great difficulties in determining chamber depth and refractive error and high myopes. Cataract surgery shows us that this can be actually quite hazardous to predict accurately.

Manifest refraction spherical equivalents were very good as 71.7 percent to 76 percent had an MRSE within .5 diopters of the target after six months and 93 percent had within target with 1 diopter at six months.

The majority of subjects gained at least one line of best corrected vision which is quite remarkable. Visual side effects, glare and halos could be expected to occur if light passes outside the limits of the lens and enters the eye through the large pupil. This should occur primarily at night when the pupil is largest.

Such issues are real but of lesser concern since many of the subjects already experienced such visual symptoms without the lens in place. Severe glare was noted at one percent at all post-op visits. Halos were more common and were moderately severe in 17 percent and severe in 3.5 percent.

Regarding the assessment and recommendations, question 1 and 2, it is my opinion that the lens is not safe for the currently intended subject population. Endothelial cell loss is a progressive problem. The damage from ongoing cell loss could be partially ameliorated by requiring the pre-op cell count of greater than 2,400, or perhaps some other number.

This would not completely solve the problem but it would help. It would also help to limit the age of the subjects. Those between 21 and 50 have different needs and issues compared with the older group.

It would be wise to be the most stringent with the younger group. The reviewer believes this lens is not safe to implant in subjects under the age of 35 regardless of the cell count. For those between 35 and 50 a cell count of at least 2,400 should be required. This would delay onset of the mean risk point, an ECD of 800 to age 75. Keeping in mind the wide 95 percent competence interval and the large standard deviations revealed in the data, this seems a reasonable level of risk.

As an alternative or additional method to reduce risk, the reviewer recommends the panel consider limiting the lens to those most in need, the group with a refraction of 9 diopters or greater. For this subset the alternatives are very limited and the added risk of late complications may be more reasonable.

For subjects over the age of 50 the late complications, 30 and 40 years away, are less threatening even though there was a real probability that they will live -- these subjects will live into their 90s. For this group a pre-op ECD of 2,000 will still probably lead to failure in 30 years. This is, nevertheless, a reasonable risk that is in line with clear lens extraction or with early cataract removal, two likely alternatives.

There seems to be a very compelling reason to limit the lens to those with an anterior chamber depth greater than 3.2. For an anterior chamber depth less than this, endothelial cell loss was twice as high and clearly unacceptable at any age or ECD. I believe it is reasonable that the lens diameter should be limited to the size of the dark-adapted pupil, although I understand that the correlation of halos and glare is not very good and has some other considerations.

That concludes my remarks. Thank you.

DR. WEISS: Thank you, Dr. Mathers.

Dr. Schein.

DR. SCHEIN: If I can make this work, I'm just going to present an overview of the comments that I submitted several weeks ago. I'll try to move quickly through anything that's been pretty well covered already. I'm purposely not going to address the individual questions at the end but to make some more general comments that I had in reviewing the protocol.

I had some frustrations in reviewing it because I felt that the work was all there but I couldn't quite extract it in the way that I needed to in order to make the assessments regarding safety that I was trying to.

First let me make a few general comments. I believe there is consensus that some follow-up of reasonable length is needed to determine safety. In the cohort I examined, we only have three-year data on about a third of eyes. It makes it hard to think of complications rates after that distance. It's greater at two years, I understand, but perhaps only about 60 percent at that time. I'm not going to get into protocol violations since we discussed that a lot this morning.

There has been a repeated theme which I would like to emphasize. When we're looking at safety, I would like to know safety not in some subgroup of patients, this Group A. I would like to know safety across the entire cohort that underwent the implantation of the device.

Obviously it would not report efficacy in a group, particularly efficacy related to corrected acuity in individuals who didn't meet a standardized entry acuity level but I do want to know this for adverse events.

There were, for example, about 50 eyes which were excluded from that primary analysis of Group A who appeared to have about twice the adverse event rate as defined by the sponsor. Likewise, I would like to look at safety issues or complication rates that in some way reflect the duration of time in the study.

For subjects that are lost to follow-up, there was a table which I've referenced there where about 20 percent of them have some worrying anatomical or functional feature noted on the last exam recorded. Of course, these are patients that are excluded from the two or three-year rates of complications.

Another issue is I would like to see adverse events and safety talked about presented not just on a per-eye basis but on a patient basis. Certainly a patient who has a retinal detachment on one eye would view the procedure as risky even in the presence of one eye that didn't have such a problem.

The intent of this device is as a bilateral device and ultimately it will be used almost exclusively as a bilateral treatment much as contact lenses are used. So similarly at different places in the report there are different rates that were given. A quoted a rate of 3.4 percent, again, is not on a person level. It's on an eye level.

It's not accounting for variable length of follow-up so in these three year cumulative rates where a denominator of 662 is quoted, I don't know how to -- I don't know what inference to draw from that when I have less than one-half the potential data at three years in hand.

There are references throughout the PMA and in the proposed labeling for comparisons with anterior chamber intraocular lenses. I think there may be historical reasons why these are in the document but I think they are inappropriate comparisons since patients undergoing anterior chamber intraocular lens are typically older. They are often already aphakic or they are in the process of suffering complications from cataract surgery, not a good comparison to make.

Looking at the safety issues, again, trying to figure out what the rates were, I had more difficulty trying to understand the differences between what were termed complications and/or adverse events. Lens opacity was listed as a complication but not cataract extraction. That seemed to be listed under other procedures. I found a couple under lens exchange.

The resuturing of a wound leak in the early postoperative period was called a secondary procedure. In my cataract practice I would call that a complication. It's not the same as a secondary refractive procedure downstream to make more accurate the efficacy of the procedure. So there's inconsistency. Retinal detachment is a complication not listed as a secondary procedure. Again, all presented on a per-eye basis alone.

I would propose that complications be divided in analyses into those which have clinical significance with an obvious potential to cause harm and I've labeled a few of them here. There are others. And to distinguish those from I would call more trivial events such as the need for punctual occlusion or the need to widen a peripheral iridotomy.

The labeling of activities like needed to resuture or reposition an intraocular lens as a nonadverse event makes no sense to me clinically. Again, frustrating in trying to figure out what the true rates of adverse events actually was. Let's try to separate the things of clinical importance from those which are not.

Similar, this issue of something parsing events that might be potentially avoidable versus not. I have even more trouble since I see no way to divorce the device itself from the surgical procedure that accompanies it. The material is polymethymethacrylate and is fine and inert. It has a wonderful track record but you cannot separate the two of them.

There is a presumption that the device and retinal detachment or, for that matter, development of cataract may be unrelated and that these are high myopes who are going to get these complications anyway. In the absence of a control group I think the sponsor takes the risk of a presumption of exactly the opposite.

The enrolled cohort here appropriately could not have had retinal detachment in the eye that was being enrolled either in the past year or in the past decade except for patients that were 30, 40, and 50 years old. By definition having not had them even though they were at risk, this is a group that is in a sense a survivorship group whose anticipated rate of such adverse events over a one, two, or three-year period would be expected to be lower, not higher.

So lens opacities, I believe, were recorded in about five percent of eyes but in the absence of a standardized grading system. I think someone made the implication earlier that in 1997 there was no concern about an aphakic intraocular lens in development of cataract and no understanding that endothelial cell counts were problematic and required multiple testing, repeat testing no matter what the name of the device was. I reject both of those notions. These things were well known in 1997.

It's difficult to assess. I don't know whether it's five percent or one percent or 10 percent. I am more concerned actually with the time dependent rate of cataract development than I currently am with projections three decades down stream for endothelial cell loss because as these patients age, they are likely to develop cataract, particularly if there is a risk of this device.

Such patients will have difficulties measuring the intraocular lens to replace and these patients will undergo cataract surgery combined with an anterior chamber lens removal which will certainly add to the risk of cataract surgery.

Regarding patient-reported visual side effects depending on one's perspective you could look at this either as an efficacy or a safety issue. In looking at it from the safety perspective, I focus on individuals who do not report the symptom before surgery and then develop it later.

It's very nice that there are individuals who report it before who do not have it later. Again, from a safety public health perspective this is the group I'm most interested in and 15 to 30 percent developed symptoms of varying severity, usually not too severe but were ones that were not noted preoperatively.

This is something that I think can benefit from further analyses to see whether there were subgroups, age, gender, degree of refractive error, the obvious kinds of parameters to see if there are subgroups with really, really large rates that would be part of a patient education or even a labeling issue.

Finally, endothelial cell counts were left unfortunately with having to draw inferences from data sets each one of which, I think, has substantial problems and limitations. Unfortunately, each data set does not compliment the other, at least in a meaningful way that I can see.

I'm actually drawn most towards the full data set. Although the image quality is poor, there is no reason to think there is a systematic bias towards under or over reporting. About 25 percent of individuals seem to have lost 10 percent or more cells which was substantially more than the proportion gaining 10 percent or more cells. I can't recall.

I think it was in the range of three to five percent that gained. So if it were purely noise, I would expect an equal distribution. Again, I couldn't tell from my own review how individuals who had secondary procedures or problems were handled or whether they were included or excluded.

This we've discussed further. Reanalyzing data reduces the individual variability but, as Dr. Stulting just point out, the measurements are still problematic because of test, retest or interpretation and reinterpretation variability.

So we are left with non-US data. The Canadian data is means only and I feel very strongly that looking at means is not the way to look at endothelial cell count again. From a safety perspective we're interested in the worse X percent. We can argue whether it's five or 10 or 15 percent or 20 percent or more cell loss but it's that part of the distribution that you're worried about from the safety perspective.

European data has all the problems that we've already discussed. A third of the patients had lost 20 percent or more cells by 10 years but, again, I don't know how much faith to put in such a small sample. I think it would be worth some discussion to get some consensus on how much of incremental loss would be of clinical significance.

We've talked about 1,200 being a floor but I reject having a rigid final cutoff because of the anticipation that a large number of these patients are likely to undergo cataract surgery and lose another five to 25 percent based on that last intervention.

To summarize, I have concerns based on the data that's presented to date that is incomplete in comparison to what will presumably be collected over the next 18 months. Additional analyses of the kinds I've recommended and the three-year data, in other words, without any new data collection on patients that haven't been recruited, I think, would go a long way.

Particularly to these nonendothelial cell count issues one would be able to see whether the rates of retinal detachment and cataract surgery, lens reposition opacities was actually on the increase or whether there were things that tended to occur early and then flattened out. That would be very important to know. Thank you.

DR. WEISS: Thank. Dr. Macsai.

DR. MACSAI: Before I start, I would like to acknowledge -- I would like to thank the Agency for this opportunity to review this PMA. I would like to acknowledge the sponsor's work in putting it together and the extraordinary analysis by Drs. Lepri, Gray and Calogero.

In addition, I would like to echo some of Dr. Schein's sentiments. This was a very difficult PMA to analyze. It was difficult for a number of reasons but they mostly have to do with lack of standardization and probably protocol design.

As I said earlier, I think that we need to look at this in light of what we know about anterior chamber IOLs and what are the risks of phakic IOLs wherever they reside within the eye.

I submitted to the panel and to the Agency a long primary review which I know the sponsors received so what I would like to do is just highlight a few issues that I think warrant our review. The first is that of accountability. I felt the accountability of this PMA was moderate.

It dropped below 75 percent at the three-year exam. Dr. Stulting did tell us patients were only told they would need to be enrolled for two years. But what is of concern is that 53 percent of the subjects in this study are ongoing and perhaps we are looking at an incomplete data set.

Eleven percent were discontinued and of these that were discontinued we learned earlier some were lost to follow-up and others had problems with the device. Those that had problems with the device are inappropriately grouped as discontinued. They should be listed as complications or treatment failures.

Enrollment. Of the 684 subjects 184 subjects were enrolled with protocol deviations in one or both eyes. This was discussed and apparently the Agency cleared these but, in my opinion, this is an alarming number of patients with protocol deviations. If the protocol is set up by the sponsors, perhaps they were too rigid in their initial establishment of enrollment criteria.

If you look at it this way, 25 percent of the subjects do not meet the enrollment criteria and this is making it even more difficult for us to analyze both the safety and efficacy of this device. When we look at criteria for safety and efficacy to quote, "The rates of cumulative and persistent complications should not exceed those of the FDA grid for anterior chamber IOLs."

I know this has been mentioned before but I have to go on record as saying this is not acceptable to me. The safety criteria for a phakic IOL should not be compared to that used during cataract surgery for an anterior chamber intraocular lens. In 2004, 1998, 1990 you used an anterior chamber IOL because things had gone wrong, disastrously wrong during cataract surgery.

In those patients an anterior chamber IOL was a second choice. Why would we compare an elective procedure that's refractive to an acceptable grid for a second choice in the treatment of a pathologic condition? The phakic IOLs must be held to a much higher standard than that of the FDA grid for ACIOLs. If it is acceptable to some to make this comparison, then we have to look at the historical perspective of what has happened with ACIOLs in the United States and what has happened with numerous ACIOL designs, their effects on endothelial cells, the fact that most of the cornea surgeons at this panel meeting cut their teeth doing transplants and removing these anterior chamber IOLs.

We knew long ago about the risks of endothelial damage with the anterior chamber IOLs. If we are going to set this PMA up as comparable to the FDA grid for ACIOLs, then I think we should also be very careful about saying that we in 1997 did not necessarily have knowledge of endothelial cell standardization or damage, etc.

Dr. Stulting has gratefully produced some information about Group E which is the eyes not included in Groups A and B in which this was used compassionate use or custom made lenses or eyes that did not have a best corrected vision of 20/40. This data really needs to be reported to the implanting surgeon and the consumers.

It's a very, very important safety criteria. It's critical to know what happens when this lens is placed, for example, under a transplant or if it's a custom designed implants. The consumer must have this information and the information needs to be segregated based on the power of the IOLs, the age of the patients, the reason that the patients are in Group E.

Lens opacities. Twenty-six of the eyes had preoperative lens opacities. I said this at the beginning, they were not measured in any standardized manner. If you can't measure them standardized preoperatively, you can't measure them postoperatively and I think a comparison is ludicrous.

You're comparing apples to oranges. What's my opinion is different than your opinion as far as cataract formation in a lens. This is not able to be scientifically evaluated with this lack of standardization.

What about the safety of all lens powers? Well, only three implants were placed under 7 diopters. This is a very small end allowing for absolutely no statistical significance. What about the role of corneal abnormalities? It was very hard for me to figure out from this PMA what was defined as a corneal abnormality.

Was it Fuch's dystrophy or was it a little foreign body scar from contact lens wear? I don't know. Without that knowledge I can't tell if there is a skew in the endothelial cell count data that may result from including these 41 eyes.

Adverse events. The sponsor stated that they thought an adverse event of one percent was acceptable and here I will echo the comments of Dr. Schein. You cannot arbitrarily decide what an adverse event is. Anything that happens as a result of the procedure that's bad is an adverse event.

If you look at these numbers of retinal detachment, cataract lens haptic dislocation, power calculation errors, inflammatory response, lenses explanted, lenses exchanged, lenses reattachment and surgical trauma, the numbers are much higher.

It's about a 3.9 percent incidence and I think that's per eye. I'm not sure if it's per patient. I really couldn't tell from looking at the data and I think it's really important to the consumer that they know the difference there because if they see it's per eye and they have two eyes, they may say, "Gee, is it twice that," whether we know or not the statistical validity of that assumption.

Patient symptoms. Again, it's very nice that they segregated out for us those patients who preoperatively responded no and postoperatively responded yes. This removes the confounding variables of glare problems that we know are prevalent in this highly myopic population. It is very significant that in patients with pupils over 5.5 mm under mesopic conditions halos were reported in 23.8 percent. These are very high numbers.

These are very high numbers because the sponsor took the time to segregate out the pre-op response being no and the post-op response being yes. In many studies this has not been done so this is basically induced problems either from the procedure or the device or the surgical technique but they are induced problems.

Pressure. I have to defer to Dr. Coleman. I'm not a glaucoma specialist. Unfortunately I'm not good at even maybe defining it as Dr. Stulting alluded to those of us who are cornea surgeons, but I was very alarmed that gonioscopy was not performed in any of these patients preoperatively or postoperatively.

I am very concerned that in the darkly pigmented patient the role of pigment dispersion from this lens may be very high. We just don't know yet but I find it hard to imagine that enclavation of the iris would not result in pigment release, flare, some level of chronic inflammation, and possible acceleration of cataract formation or glaucoma.

The endothelial cell data was very difficult to analyze. It's been adequately, I think, addressed by Dr. Gray, Dr. Mathers, and Dr. Schein. I have very little new to add. You all know that from baseline to three years the decrease was 4.7 percent but this loss seemed to be higher between the second and third year as compared to between the first and second year intimating that there is an increase in endothelial cell loss over time taking into account the lack of standardization of what we're looking at.

Anterior chamber depth was addressed by Dr. Mathers. There were only six eyes but in those eyes there was an alarmingly high rate of loss of endothelial cells alluding to the fact that the depth of the anterior chamber plays a big role in endothelial cell loss in these patients either from surgical trauma or ongoing trauma from eye rubbing or something of that sort.

The number of eyes that demonstrated greater than 10 percent loss was analyzed and it was looked at. I don't need to go on about this. And a consistent cohort was also looked at showing 2.38 percent overall loss. Just clearly the endothelial cell count has not stabilized in this short time period that we're looking at during this accelerated review.

I don't know what the endothelial cell loss rate is but it's somewhere between 1.58 and 3 percent. I think that 2,000 cells per millimeter-squared is way too low of a cutoff, especially in a 21-year-old.

In summary I'll tell you that I had a very hard time reviewing this PMA due to lack of standardization and enrollment criteria, outcomes reporting, lens characterization, adverse event definition, gonioscopy, and specular microscopy. And though I'm not sure, I've chosen purposely not to answer the panel's questions during this presentation. I would use this time to ask the Agency and sponsors who are in the development process of improving our field by creating these phakic IOLs that it's very difficult to give a fair and reasonable analysis of safety and efficacy without standardization of these key features. Thank you.

DR. WEISS: Thank you very much. I want to thank all of the reviewers for their excellent and clear presentations. At this point we are going to go on to the panel discussion of this PMA. I would ask FDA if they could come forward to the podium and then just present each question so that we can discuss it in order.

While Dr. Lepri is doing that, the first question which I will just read out is, "Do the endothelial cell data presented in the overall analysis stratified by anterior chamber depth and the extrapolations over time provide reasonable assurance of safety in the ARTISAN myopia lens."

What I would like to do is just go around and get the opinions. If you want to give me a yes or a no, that's the best opinion possible. If you want to add some comments, that's okay as well.

Dr. Schein, do you think that there is reasonable assurance of safety on the basis of the endothelial cell data, question No. 1?

DR. SCHEIN: No.

DR. WEISS: Dr. Bandeen-Roche.

DR. BANDEEN-ROCHE: No.

DR. WEISS: Dr. McMahon.

DR. McMAHON: No.

DR. WEISS: Dr. Bradley.

DR. BRADLEY: I think it's impossible to project out 30 years. My answer is I don't know.

DR. WEISS: I don't know. Okay. Dr. Macsai.

DR. MACSAI: From the analysis of what I was given to review, I would have to say no.

DR. WEISS: Dr. Grimmett.

DR. GRIMMETT: In short, no.

DR. WEISS: Dr. Mathers.

DR. MATHERS: No, but I do think that the age of the patient when this is performed plays a role in that decision.

DR. WEISS: Dr. Casey.

DR. CASEY: No.

DR. WEISS: Dr. Coleman.

DR. COLEMAN: No.

DR. WEISS: Dr. Van Meter.

DR. VAN METER: No. This may not be the time to discuss it but I think that it's reasonable to talk about whether or not we want to lump all surgical and operative issues in with the device itself because we ourselves have said that anterior chamber lenses for pseudophakic correction are not a legitimate comparison because of the differences in surgical technique. These are sick eyes and they've had previous surgeons.

DR. WEISS: Actually, since we're not -- I just want to speak to the particular question so that may --

DR. VAN METER: No.

DR. WEISS: Dr. Smith.

DR. SMITH: No.

DR. WEISS: Dr. Huang.

DR. HUANG: I don't know.

DR. WEISS: Is there anyone that requires any discussion on this point? I say this with great hesitancy. Is there anyone who just requires some discussion? Personally I think many of the points, if not all the points that are relevant, have already been elucidated. Okay.

Dr. Lepri, do you need anymore information from the panel on Question No. 1?

DR. LEPRI: I would say no. That was pretty clear cut to me.

DR. WEISS: We're trying. Question No. 2. I think this way of going around the table does work so we're going to try this another time. Question No. 2. "Do the other data, not the endothelial cell data but everything else, presented in the PMA provide reasonable assurance of safety?"

Dr. Schein.

DR. SCHEIN: No.

DR. WEISS: Dr. Bandeen-Roche.

DR. BANDEEN-ROCHE: No, and I would just like to second Dr. Schein's concerns about having to take into account time under observation for incidence of events.

DR. WEISS: Actually, from my elucidation, I would -- you can contradict me if you like. Would it be helpful to you if whoever -- if someone feels that the other data do not provide reasonable assurance safety, if they just specify what data they are concerned about?

DR. LEPRI: Exactly. I was just going to mention that to you. If you specify what in particular made you make that decision, it would be helpful to us.

DR. WEISS: So, Dr. Schein, you felt the other data do not provide reasonable assurance of safety. Can you just elucidate what your particular concerns are?

DR. SCHEIN: Lens opacities, retinal detachment. Need to move, reposition, or exchange the implant.

DR. WEISS: Are you concerned that there's a higher rate of retinal detachment with this lens than the normal patient?

DR. SCHEIN: The concern is that the procedure coupled with the device adds significant risk of retinal detachment compared to not having the procedure or device.

DR. WEISS: Dr. Bandeen-Roche.

DR. BANDEEN-ROCHE: Yes. I would rely on the clinical expertise to specify where there's a concern. Then I just felt like the incidence rates that we've been given are probably undercut because they are not presented in a Kaplan-Meier or taking into account time under observation.

DR. WEISS: So you had safety concerns because the statistics as they were presented didn't give you the information you wanted?

DR. BANDEEN-ROCHE: Yes, in combination with the clinical concerns expressed by my colleagues.

DR. WEISS: Okay. Dr. McMahon.

DR. McMAHON: In the aggregate, no. If you look at the complication or adverse event rate as compiled by Dr. Schein and Dr. Macsai, the incidence rate is too high. If you start talking about individual rates, I have a hard time getting a handle around it to know whether that is too high individually or not.

DR. WEISS: So, from what I understand that you're saying, it's hard to answer this question because you don't have the numbers that you want.

DR. McMAHON: Correct.

DR. WEISS: What numbers would you want from sponsor? What would you like to look at which would allow you to make that determination?

DR. McMAHON: I think the time dependent issues that have already been raised are the ones that I would be looking for.

DR. WEISS: Dr. Schein.

DR. SCHEIN: And the parsing of events and complications from those with clinical significance separated from those without.

DR. WEISS: So basically put all the adverse events together and also put them in a format so that it's per patient and not per eye.

DR. SCHEIN: Or both.

DR. WEISS: Both. Anything else in terms of the statistical? Any other things that I have not mentioned that you would want?

DR. SCHEIN: No. I think Dr. Bandeen-Roche emphasized we want the amount of time or timeline.

DR. WEISS: We want a timeline. We want binocular. We want monocular.

If you could just speak into the microphone so we can hear. Could you just repeat that so we can make sure we got it on the transcript.

DR. SCHEIN: I don't know how far back to rewind.

DR. WEISS: Tell us your wish list.