Скачать 330.02 Kb.

DR. STARK: I think that's a good point. Many of these people are doing corneal modeling after surgery. They do it when the epithelium clears up at 2 or 3 months and then later. And so far, there are only these three cases reported. Those are the only cases in the literature. Richard Abbott's case is after three LASIK; that's just bad surgical mentality, I guess. Those are the only three reported cases, but they are of concern, because they are reported by an experienced investigator, Seiler, who I believe, from reading the article I can't remember for sure but I believe they did the cases, so you would expect that they had ruled out keratoconus prior to the surgery. DR. McCULLEY: We also don't know how many cases have been done below the 250, with the 250 being the guideline. It seems like until we have a series with significant numbers of patients measuring the thickness of the flap, knowing how accurate that is, and therefore exactly how much posterior cornea is left undisturbed, until we have that data, I don't see how we can get off of the 250. Does anybody else? Mike? DR. BELIN: I was just going to agree with Mark. We actually don't know if we get ectasias, because what we are doing is we are using topography, and we are doing pre and post, but we have flatted because we are dealing with higher corrections in LASIK we are flattening the central cornea. You would either have to look at patients that you've done cuts and no ablation, which is never going to happen, or you look at the peripheral cornea, and the problem with that is we know the reliability of the topography out there is very, very poor. So it really is an unanswered questions. Part of the reason we may see regression, which we are, people admitting in the higher corrections, is we may be getting a slight ectatic effect. You don't really know. DR. McCULLEY: Other comments? [No response.] DR. McCULLEY: Do you have any further questions? [No response.] DR. McCULLEY: We'll allot 10 minutes for comments from anyone in the audience who has an additive view or a differing view. [No response.] DR. McCULLEY: Seeing none, Morris, do you have anything further? DR. WAXLER: I have one more question that we had mentioned at the beginning that you have nothing written on, and I think it is an issue that has come up repeatedly, and I will address it if this is an appropriate time. DR. McCULLEY: Certainly. DR. WAXLER: Currently, the guidance defines refractive stability as 95 percent of the eyes reaching a change of less than or equal to 1.0 diopter of manifest spherical equivalent refraction between two refractions performed at least 3 months apart. The guidance also suggests confirmation of the stability estimate with another measure of MRSE 3 months later. There are two aspects of refractive stability that we would like you to discuss and make a recommendation to the agency on: How much change in refraction is reasonable to expect? A mean difference of zero refraction over a 3 month interval is ideal, but what is a reasonable change to expect 0.01 D, 0.10 D, 0.20 D, 0.30 D, 0.50 D? The current guidance implies that the change can be +/ 1.0 D. Is this reasonable? Should we add a statistical definition so that the slope of change in refraction should not be significantly different than zero? That's the first question; it was a long winded one, but that's the first question. Do you want to deal with that one and then go to the next one? DR. McCULLEY: Why don't you let us know what you have in store for us? DR. WAXLER: Okay. The second one is if 95 percent of the eyes are within +/ 1.0 D MRSE between two refractions taken between 3 months and 6 months, then, has stability been reached at 3 months or 6 months? Should confirmation of the point of stability be taken at 9 months or earlier? DR. McCULLEY: Okay. We'll go back to your first question in just a minute. When I re read the guidance, it didn't give a time. I had read in the guidance document itself that between the two intervals, one interpretation of it was one month apart, two refractions that were the same one month apart that was the CRS interpretation of it in Dr. Kezirian's document. When I read the guidance document, it said less than 1.0 D of change between two visits. It didn't say 3 months. DR. WAXLER: Yes, it does. DR. McCULLEY: What page? DR. WAXLER: I don't have the page number with me, but I'm sure my vast auxiliary staff here will find it. DR. McCULLEY: Okay. I'll look again. DR. WAXLER: I just checked it this morning, and it said it when I read it this morning. DR. McCULLEY: Okay. I read it in a different place, and it didn't say 3 months; but here on page 18, it does. Okay. So it is stated as 3 months. Now that you've gotten me straightened out DR. WAXLER: Actually, to make that first question simpler DR. McCULLEY: Please. DR. WAXLER: you have a mean difference between those you had the 3 months the guidance only says the 3 month interval and 95 percent within +/ 1.0 D, but do you want that curve to be asyntotic [ph.], and at what point do you want it to be asyntotic? You know, if it changing at 0.3 D, and it is still changing, the person is going to be back where they started not too long in the future, so the question is how close to zero do we need to have that estimate. Again, it's one of these things that you don't like to deal with, but we have to make a judgment when an application comes as to whether or not it is asyntotic, and the change is not unacceptable from a clinical point of view. DR. McCULLEY: Okay. You said something about mean. This is 95 percent of individual patients have less than 1.0 D of change between two visits 3 months apart. DR. WAXLER: Right. DR. McCULLEY: The only way we can judge that curve other than those two points is to have other points to know what the slope of that curve is, or what the history of it is apt to be. So I don't know that we can answer that from two points. DR. ROSENTHAL: But you have all the points. You have the mean difference. Not only do you collect the percentage that has less than or equal to 1.0 D between the two, but you have the mean difference between the DR. McCULLEY: The whole group. DR. ROSENTHAL: no the mean difference between the refractions at the two intervals. DR. McCULLEY: Right, and we're saying that 95 percent of the people have to be within 1.0 D difference. DR. ROSENTHAL: That's right, but when you look at the means, they can DR. McCULLEY: What means? We're looking at individual patients in this. DR. ROSENTHAL: The other way to look at it is you take a mean DR. McCULLEY: Of the total group. Two different issues. DR. ROSENTHAL: a mean of the differences DR. WAXLER: The mean change. DR. ROSENTHAL: The mean change. Sorry. Thank you. A mean change, a mean change. That's what I meant to say. DR. McCULLEY: Okay. DR. ROSENTHAL: And the mean change should get smaller and smaller as you get closer to stability. Now, if the mean change of .3 we know if it is .01, it is quite acceptable, or if it is .05, it is quite acceptable, or even if it is .1, it is probably acceptable, but DR. McCULLEY: The mean change is for the total population. DR. ROSENTHAL: Correct. DR. McCULLEY: And we get that data. DR. ROSENTHAL: Yes. DR. McCULLEY: In a different presentation of the data. So you are saying that to DR. ROSENTHAL: In the presentation of stability, we get the issue you first dealt with, and we also get the mean difference. DR. McCULLEY: The mean change; right. And do we have a number for mean change? DR. ROSENTHAL: No. DR. McCULLEY: So you're asking should we add to stability in addition to 95 percent of the patients being within +/ 1.0 D or, within 1.0 D 3 months apart that we see if we can come up with an acceptable mean change between two time points to predict stability, which would also presumably have to coincide with the 95 percent being within a diopter change. DR. ROSENTHAL: Correct. DR. McCULLEY: Now I understand. Dr. Belin? DR. BELIN: Does it make sense for anyone that we have a definition of preoperative stability that's so different from the definition of postoperative stability? The definition preoperatively is 0.50 D or less during the year prior to baseline exam; the definition postoperatively is up to 1.0 D over 3 months. To me, that doesn't make any sense. Now, I don't want to make anyone do a year long to show stability, but I think 1.0 D over 3 months is not stable. DR. MACSAI: Hear, hear. Second. I also would point out that I know the agency has access to it, and I didn't know you were going to ask this, so I didn't bring any of the reprints but earlier, someone referred to Carla Zadnik's study on the reproducibility of refractions, and we do know there is some change in manifest refraction whether or not you have refractive surgery over time. So that perhaps a slope of zero may not be within the nature of the human beast, but perhaps from those two sources, you could calculate what an appropriate slope would be. And I would also agree with Dr. Belin's comment that to have a preoperative acceptability of 0.50 D and postoperative of 1.0 D is inappropriate. DR. McCULLEY: Now are you sorry? What is the number I know I have in my mind what the number is reproducibility of a refraction, the spherical component over time? Dr. Bullimore? DR. BULLIMORE: From memory, the 95 percent confidence intervals from Carla's paper were manifest subjective refraction with the order, I think, of +/ 0.50 to 0.75 but that's DR. BELIN: But for the average population, that doesn't take into that's an individual variation over time. The mean population is stable. DR. BULLIMORE: Yes. As was raised from the floor earlier, we have got to consider two issues here the repeatability of our measurement and whether there is an underlying change in refraction, regression or whatever you want to call it creep and whether there is any inherent variability induced by the procedure the patient has undergone. So we are trying to cram all these different effects into a nice, bite size number that Morris can go away and take with him, and it's difficult. DR. McCULLEY: Marian? DR. MACSAI: This is a point at which postmarket surveillance could answer the question to some degree. I recall in previous discussions about individual PMAs that we are not discussing that this issue has come up over and over again. The postmarket surveillance was done in the PERK study by looking at the patients 10 years out, and they found some interesting findings. DR. McCULLEY: Other comments? Morris? DR. WAXLER: If I could just add to what Marian said, I think that our immediate issue, though, is in the premarket, because we have applicants with particular data, but we want to be fair, and we want also to do a good job from a public health standpoint what is a reasonable value from the standpoint of a group of clinicians. If you had a group of patients, and you had a certain change of 0.2 or 0.5, would you be comfortable with a laser that had that kind of mean difference DR. McCULLEY: Mean change between the two points of stability as otherwise defined by our somewhat flawed numbers. DR. WAXLER: Right. Otherwise what happens is we get into a game where we're saying, well, no, this one looks good, and that one looks bad, and you hardly have a stable footing. DR. McCULLEY: Yes. It would have been nice considering the way we approached these other questions, with one person taking the assignment and really searching the literature, we could probably do a better job for you. DR. WAXLER: Yes, I think so. DR. McCULLEY: I don't get a sense we are heading in any strong direction. Does anyone else have anything? Dr. Wang? DR. WANG: Ming Wang. With regard to the duration postop, I think there are some clinical observations which would support one year. In terms of wound healing, the flap is almost impossible to lift at one year, meaning the stroma really healed quite well. Second, most of the refractive surgeons in clinical practice do not follow patients beyond one year, and that seems to be working very well. Third, to second Michael's point, to have equal standards postop, since preop, we have stability of refraction one year, say, 0.50 D, it would be reasonable to have. So these three all support intuitively to me one year follow up. DR. McCULLEY: Number one, I can raise flaps at a year. The interface probably never heals. And I have lifted flaps at a year. it may be a little bit harder than at 3 months, but I don't think the interface probably ever heals. And from maybe an ideal standpoint, to require a year or two or even longer, the more comfortable we get but again, we start to get into the issue that I mentioned before, that we start to delve over from ideal science into the world of commerce, and there are some things where we have to make some compromises in the scientific ideal and the practical. Any other comments? Dr. Belin? DR. BELIN: I made a point of suggesting that I don't think we should have one year; I agree with you that that is not practical. But 1.0 D over 3 months and again, this is not going to be just for lasers; we have some devices, instruments or implants that may be applicable for only low levels of correct. Let's say we have a device that's only applicable from 1.0 to 3.0. Let's say it corrects 250, which for us is +/ 0.50, and that's a perfect result. That means that at one month, they are 250; at 3 or 4 months, they are 150, and that meets criteria, and that's stable. It's not. It meets everything that we have est up, but it's not stable. DR. McCULLEY: So you are back on the point that within 1.0 D 3 months apart is probably not a good guideline, and you would suggest it be 0.50. DR. BELIN: I'll just ask does anyone here think a 1.0 D change over 3 months represents stability. Does anybody that's all I'll ask. DR. McCULLEY: Okay. Fair question. Straw poll. I won't try to restate it. Who thinks that a 1.0 D change over 3 months or less than 1.0 D is an acceptable definition of stability anybody? Dr. Ferris? DR. FERRIS: This is Rick Ferris. It's confusing, because I think we have to be careful with this 5 percent have the 1.0 D change. If you expect 5 percent to have a 0.50 D change by chance, just measurement error, and you have a real 0.50 D change in I guess you can figure out what proportion you have to have a real 0.50 D change to then see a 5 percent 1.0 D change I guess you would almost have to have 100 percent having a 0.50 D change to get 5 percent having a measured 1.0 D change. But you can see the issue has to do with talking about individual patients, which I think is difficult, but you have to do it with regard to complications, and then talking about the group. And when you are talking about this drift over time, I think you are stuck with talking about the group. It is very hard to look at individual patients. It isn't very hard to look at the change over time, and that statistic, the average change at each visit, is a relatively simple statistic to create. And if that is, as Morris suggested, that statistic ought to be flattening. It shouldn't continue. If you had a 0.25 D average loss in the first interval and continued to have a 0.25 D average loss over the next two intervals, that would be much more concerning to me than even an initial change in the first interval and then a flattening. 