|
|
Post by jpg on Mar 30, 2014 15:36:49 GMT -5
The discrepancy re PD/PK re GIR/dosing is a concern to me as well. Putting my wishful thinking hat on, I imagine FDA could leave this as a fine-tuning exercise between physician and patient, especially given the observed individual variability. Putting my worse case hat on... never mind. You would think that between the 2 CRLs the company had plenty of time to sort through and work out these known issues. Very sloppy. Grrr. The problem is they were testing a new device and if they gave to much they probably feared hypoglycemia and if they give a bit less they probably figured that dosing will simply be adjusted upwards (which it was) and that every patient is different and what counts are the glucose readings anyway. We treat to glucose levels based on rough understandings of PK/PD anyway. Who really cares about 3/10 or 4/10 in a long run clinical scenario? It is like arguing about fentanyl to morpine or hydrocortisone to solumedrol potency equivalences. A ballpark figure is necessary but past a cerain level it is not clinically relevant (not to say the FDA won't care but I doubt the panel members will). I did see 2 pharmacology students argue once if fenanyl to morphine was 1/80 or 1/100 for a surpisingly long time. They didn't ask my input but I would have laughed as it is a meaningless argument because the half lives are different with different age groups, the peak and redistibution are different and more impotantly it doesn't make any clinical difference in these ranges. The same could be said for prandials and Afrezza. 3/10 or 4/10 at what time? 10 minutes? 60 minutes? 100 minutes? 3 hours? If anyone comes up with a uniform potency answer to all these time points the are simply wrong and not thinking it through. The whole non linear thing is also kind of a theoretical argument and has to do with the same exact arguments as above and a massively complicated second set of variables that have to do with stuff we don't know a lot about yet because there has never been an insulin like Afrezza to study all this stuff. And no MNKD could not have done all these studies since we are talking of 100s of studies that will be done like was done when prandials first came out. I do know the non linear thing being said to be due to a faulty design of the Gen2 inhaler is wrong and simplistic. There will be many academic careers built around understanding a few of these topics. As another poster here or on another board pointed out: look at the original understanding (andI would add clinical interest or lack of) of the first prandials to come out. We know a lot more about Afrezza then we did about those first meal time insulins already but we will hopefully get to another level with an approval and the right partner. JPG |
|
|
|
Post by alcc on Mar 30, 2014 17:01:32 GMT -5
I do not disagree that from a clinician's perspective, PK, PD, PK v PD are of little more than theoretical interest. However, to the extent FDA must look at proper support for dosing, these become more than theoretical concerns. When in doubt they may send the NDA back with at least a partial response. Even more to the point, looking at the less than stellar results in 171, I wonder if the company had done enough internal research (plus P2 data) to come up with the best titration guidelines re both Afrezza and basal. Empirically, it does not look they did.
Curious, is a sponsor permitted to do internal "trial/mock" runs to fine tune trial protocol, like a mock trial? Or is that verboten?
|
|
|
|
Post by liane on Mar 30, 2014 17:23:46 GMT -5
That's what the Phase 1 & 2 trials are supposed to determine. These are smaller scale trials to evaluate safety and efficacy at differing doses. I haven't looked at www.clinicaltrials.gov lately to see if these were done on the DB or just MT. I think there were just the in vitro studies which MNKD was asserting showed the equivalence. And now the 171 study was to show the in vivo equivalence. You cannot sneak in a mock run trial as all studies on humans must be registered. |
|
|
|
Post by esstan2001 on Mar 30, 2014 20:05:54 GMT -5
The problem is they were testing a new device and if they gave to much they probably feared hypoglycemia and if they give a bit less they probably figured that dosing will simply be adjusted upwards (which it was) and that every patient is different and what counts are the glucose readings anyway. We treat to glucose levels based on rough understandings of PK/PD anyway. Who really cares about 3/10 or 4/10 in a long run clinical scenario? It is like arguing about fentanyl to morpine or hydrocortisone to solumedrol potency equivalences. A ballpark figure is necessary but past a cerain level it is not clinically relevant (not to say the FDA won't care but I doubt the panel members will). I did see 2 pharmacology students argue once if fenanyl to morphine was 1/80 or 1/100 for a surpisingly long time. They didn't ask my input but I would have laughed as it is a meaningless argument because the half lives are different with different age groups, the peak and redistibution are different and more impotantly it doesn't make any clinical difference in these ranges. The same could be said for prandials and Afrezza. 3/10 or 4/10 at what time? 10 minutes? 60 minutes? 100 minutes? 3 hours? If anyone comes up with a uniform potency answer to all these time points the are simply wrong and not thinking it through. The whole non linear thing is also kind of a theoretical argument and has to do with the same exact arguments as above and a massively complicated second set of variables that have to do with stuff we don't know a lot about yet because there has never been an insulin like Afrezza to study all this stuff. And no MNKD could not have done all these studies since we are talking of 100s of studies that will be done like was done when prandials first came out. I do know the non linear thing being said to be due to a faulty design of the Gen2 inhaler is wrong and simplistic. There will be many academic careers built around understanding a few of these topics. As another poster here or on another board pointed out: look at the original understanding (andI would add clinical interest or lack of) of the first prandials to come out. We know a lot more about Afrezza then we did about those first meal time insulins already but we will hopefully get to another level with an approval and the right partner. JPG |
|
|
|
Post by esstan2001 on Mar 30, 2014 20:15:19 GMT -5
jpg, alcc, liane, mnkdfan, thanks for a lot of valuable input.
jpg, is it likely that the Mannkind presenters succeed in convincing the FDA on such matters as how irrelevant this dose conversion to IA is (beyond providing a fair dosing start point) and can they get the FDA reviewers to weight the other benefits that are not covered using the standard metrics for A1C, PK, PD etc? More weighting needs to come in on the benefits to outweigh the risks, and the FDA has seemed unusually conservative and risk averse the past 4 years.
I'm pretty comfortable on T2, I'm even getting a little less agitated on T1; I just keep reviewing the dosing section and wonder how stubborn the FDA decision makers will be on this, given it is their job to highlight the concerns...
not necessarily expecting an answer, but I do welcome your opinions.
|
|
|
|
Post by jpg on Mar 30, 2014 21:46:23 GMT -5
jpg, alcc, liane, mnkdfan, thanks for a lot of valuable input. jpg, is it likely that the Mannkind presenters succeed in convincing the FDA on such matters as how irrelevant this dose conversion to IA is (beyond providing a fair dosing start point) and can they get the FDA reviewers to weight the other benefits that are not covered using the standard metrics for A1C, PK, PD etc? More weighting needs to come in on the benefits to outweigh the risks, and the FDA has seemed unusually conservative and risk averse the past 4 years. I'm pretty comfortable on T2, I'm even getting a little less agitated on T1; I just keep reviewing the dosing section and wonder how stubborn the FDA decision makers will be on this, given it is their job to highlight the concerns... not necessarily expecting an answer, but I do welcome your opinions. I can't really know or even answer that but by saying that if they insist on exact conversion dosing then one could ask how many units of insulin of Lantus are equal to how many units of any prandial insulin? We regularly make a sum total of how much insulin a patient is on per day and think this is totally logical (and at steady state it might be 1/1 over a 24hour period for sc) but not 2 hours post Lantus or 2 hours post prandial. Giving 20 units of Lantus per day in the morning will have a very different effect then giving 20 units of a prandial before breakfast once... Giving 7 units 3 times a day of a prandial might be much closer to lantus 20 daily though. The fat prandial tail becomes a basal if given often enough. As always the devil is in the detail... This gets even trickier with a shorter acting insulin and you add another layer of complexity if the peak is in a few minutes and another layer of complexity if it is inhaled and therefor has a truly different route of absorption then sc. At least if they were comparing hexameric IV to inhaled monomeric it would be a bit easier. The easiest would be inhaled monomeric with IV monomeric I guess. Companies tweak their insulins to try and make them different in profile to mimic what happens in the body but then we need to boil it all down to a simple and simplistic conversion number... Few things in biology are that simple. The FDA knows this (hopefully!). JPG |
|
|
|
Post by harshal1981 on Mar 30, 2014 23:47:16 GMT -5
Figure 7 on Page 173 of FDA BD and the table 10 on page 76 align with explanation on reasons behind withdrawals (in paragraph starting with "The Verbatim) will be sufficient to weaken the first (most conservative) option of sensitivity analysis that requires to impute 0.4 for all the patients that didn't complete study (and in that method, recalculating CIs, Afrezza fails to prove non-infiriroty as per description on 173 and table on 174pg)
The only and most conservative method in which afrezza fails to prove non-infiriority requires the algorithm to assume worst case out come for all the patients had they all stayed and contributed the data. Details on pg 76 clearly says that majority (17) dropped out of Gen2 arm due to "unwillingness to comply with study requirements). Essentially majority of them were "Rowdy" players of the pool. Fig 7 shows the similar behavior. And hence assuming all of the 44 drop outs from Gen2 arm as Missing Not At Random (MNAR) is really not a robust way of sensitivity. NOte that the NSA report mentioned at the bottom of the page 25 (which recommends not using LOCF) emphasizes the fact that when performing MNAR, it introduces even more reliability problems (and resultant probability of false negative or false positive).
I do believe that 2ned sensitivity analysis method where 5 patients are treated at MNAR because they indicated lack of efficacy is most sensible way of sensitivity analysis and under that afrezza does meet non-inferiority.
Looking at above and the fact that afrezza's role or benfit is not only to reduce HbA1c but it is really a better disease management. Even thought trial design and end points are established as per the "current commonly accepted standards". Best argument in favor of afrezza is really that it offers best possible balance of give and take between HbA1C, Hypo and other side effects.
I think my strong belief is that we will have lots of discussion around how best use this new therapy and accordingly how to structure the label. I really don't think approvability is an issue.
Type 2 is no brainer considering 102 study and 175. Again how to best utilize afrezza to better manage the Diabetes is a point of discussion.
Of course, the label will largely drive the commercial potential and to an extent reimbursement category from payers, (and hence the impact on partnership % split or takeover price point). But I don't think here, approval is in going to be under the threat.
Its almost 1am. I think after spending so many hours making sense of the BD, I am going to sleep bit more easily then last few days.
Will see you all in DC who ever is attending AdCom in person.
|
|
|
|
Post by alcc on Mar 30, 2014 23:53:08 GMT -5
Few things in biology are that simple. The FDA knows this (hopefully!). JPG Therein is the crux of the matter: (hopefully!) |
|
|
|
Post by alcc on Mar 31, 2014 0:00:42 GMT -5
Looking at above and the fact that afrezza's role or benfit is not only to reduce HbA1c but it is really a better disease management. Even thought trial design and end points are established as per the "current commonly accepted standards". Best argument in favor of afrezza is really that it offers best possible balance of give and take between HbA1C, Hypo and other side effects. Of course. That is why I was so tee'ed off that the reviewer blithely equated fewer hypos with Afrezza being less effective than comparator. |
|
