Louis
☆    

NC, USA,
2015-11-24 22:35
(3047 d 15:31 ago)

Posting: # 15668
Views: 7,476
 

 How to use 2nd passed pivotal study after 1st pivotal failed [Study As­sess­ment]

Two pilot studies (2x2) passed for AUC and Cmax, a pivotal study (4x2) failed for both AUC and Cmax (AUC and Cmax ~130%, U95%CI ~1.40). Another large scale study passed. Would like to invalidate the failed pivotal study which had significant period effects (AUC and Cmax); a significant product*period*sequence effect; significant product differences were seen in periods 1 & 3, not in 2 & 4; within sequence product differences were significant within period 1 but not in sequence 2; CV ~ 35% for both AUC and Cmax in failed study; ~ 25% in past studies. I would like to use the second large study as pivotal. How can I do this - do I need to invalidate the pivotol study.


Edit: Category changed. Please follow the Forum’s Policy. [Helmut]
ElMaestro
★★★

Denmark,
2015-11-25 00:08
(3047 d 13:58 ago)

@ Louis
Posting: # 15670
Views: 6,406
 

 How to use 2nd passed pivotal study after 1st pivotal failed

Hi Louis,

❝ Two pilot studies (2x2) passed for AUC and Cmax, a pivotal study (4x2) failed for both AUC and Cmax (AUC and Cmax ~130%, U95%CI ~1.40). Another large scale study passed. Would like to invalidate the failed pivotal study which had significant period effects (AUC and Cmax); a significant product*period*sequence effect; significant product differences were seen in periods 1 & 3, not in 2 & 4; within sequence product differences were significant within period 1 but not in sequence 2; CV ~ 35% for both AUC and Cmax in failed study; ~ 25% in past studies. I would like to use the second large study as pivotal. How can I do this - do I need to invalidate the pivotol study.


This is all a bit confusing, and some of it makes little sense to me.
Is there anything you don't like about the first pivotal, other than its result? Period effects happen occasionally and do not invalidate results. You mention significant product*period*sequence, how did that find its way into your analysis? I have no idea what to do with such info. Out of curiosity: Did you test that against the residual or against the between-subject MS?
Do you think you just had too low sample size for the observed CV in your failed trial? Let us hear the whole story.
You might wish to audit your trials, to find out if one trial is more credible than the other. But be fair about it; it might be that the first trial turns out to be the more credible. 35% CV in one trial and 25% in the next is perhaps an OK trigger, especially if the trials were conducted under the same conditions. More details please :-)

Pass or fail!
ElMaestro
Mahmoud
★    

Jordan,
2015-11-25 09:22
(3047 d 04:44 ago)

@ Louis
Posting: # 15671
Views: 6,352
 

 How to use 2nd passed pivotal study after 1st pivotal failed

Dear Sir

❝ […] a pivotal study (4x2) failed for both AUC and Cmax (AUC and Cmax ~130%, U95%CI ~1.40).


What did you mean about 4x2 design. Is (TR,RT,RR,TT)
M.abdel Mohsen


Edit: Full quote removed. Please delete everything from the text of the original poster which is not necessary in understanding your answer; see also this post! [Helmut]
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-11-25 20:58
(3046 d 17:08 ago)

@ Mahmoud
Posting: # 15674
Views: 6,281
 

 4-periods

Hi Mahmoud,

❝ What did you mean about 4x2 design. Is (TR,RT,RR,TT)


Please read Louis’ post again. He wrote about four periods, right? Hence, it was one of the fully replicated 4-period 2-sequence designs (RTRT|TRTR or RRTT|TTRR). Nobody with a lucid mind would use Balaam’s design due to its poor power characteristics.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-11-25 20:53
(3046 d 17:13 ago)

@ Louis
Posting: # 15673
Views: 6,529
 

 More information, please

[image]Hi Louis,

like ElMaestro I’m a little bit confused. Without the additional information he requested we would have to fall back to our crystal ball…
Some remarks/questions from my side:

❝ Two pilot studies (2x2) passed for AUC and Cmax […]



Since both passed, why didn’t you stop and submit them as pivotal evidence of BE? Until very recently the FDA was happy with that (given the usual :blahblah:, like appropriate design and execution, at least 12 sub­jects). Inter­esting that the FDA removed the guidance from its website (last archived [image] 2015-02-26) and left us alone with a draft where such a statement is missing.

Why did you opt for a full replicate design in the pivotal(s)? You wrote that the CV was

❝ ~ 25% in past studies

(and you passed in two 2×2 crossovers). Were you aiming at RSABE as a kind of “safety net”? You have to sub­mit all studies’ data to the FDA. Which impression might a reviewer get? If I play the devil’s advocate:

“They had moderate variability in two pilot studies. Both showed BE. Now they want to convince us (me!) that the drug / drug product is highly variable. Maybe they opted for a cheaper (bad?) CRO in order to cut costs. I know the equation: Bad CRO = lousy study conduct = higher vari­abi­lity. In order to be on the ‘safe side’ they want us (ha, me‼) to accept RSABE. I don’t buy it.”

(The grumpy reviewer grabs the Rejected-stamp from his desk and seals the application)

❝ […] a pivotal study (4x2) failed for both AUC and Cmax (AUC and Cmax ~130%, U95%CI ~1.40).


Oops. Terrible. Which were the PEs (CIs, samples sizes) in the two pilot studies? Did you use different batch(es) of reference and/or test?

❝ CV ~ 35% for both AUC and Cmax in failed study


OK, let’s see: 90% CI ~1.21–1.40. With ~35% CV that would translate into a sample size of 62 subjects in a 4-period full replicate. You could apply RSABE (sw >0.294), but failed on the PE (above 1.25). Is my guess about the sample size correct? If yes, why was it that high?

❝ Another large scale study passed.


Fine.

❝ Would like to invalidate the failed pivotal study which had significant period effects (AUC and Cmax); […]



Irrelevant in crossovers (unless sequences are extremely [sic] unbalanced).

❝ […] a significant product*period*sequence effect; […]



Here I’m with ElMaestro. Where does this term come from – and what was the intended purpose? The FDA is pretty clear about preferred models for replicate studies.

❝ […] significant product differences were seen in periods 1 & 3, not in 2 & 4; within sequence product differences were significant within period 1 but not in sequence 2;


I would say your model is a bit over-specified. On another note, significant treatment effects are irrelevant in BE. Generally they pop up if the sample size is high for the given variability (the magic “overpowered” studies). Again: Which was your sample size?

❝ CV ~ 35% for both AUC and Cmax in failed study; ~ 25% in past studies.


Shit happens. If you tell us the sample sizes (:-D), we could calculate the respective CIs of the CVs. I bet that they overlap.

❝ I would like to use the second large study as pivotal. How can I do this - do I need to invalidate the pivotol study.


   The combination of some data and an aching desire
for an answer does not ensure that a reasonable answer
can be extracted from a given body of data.
     John W. Tukey


By “invalidating” a study you might shoot yourself in the foot – unless you find clear reasons (clinics, bio­analytics). Statistics alone is not sufficient.

The FDA in the past was fine if a failed study was repeated due to low power. Here we have a tricky case. The CV was not the problem (would have be taken care by RSABE), but the PE. Difficult to predict whether the agency will accept your justification along these lines: “Two pilots showed BE. In the pivotal the PE was 1.30. Might have been pure chance. We repeated it.” Duno.
Across the pond (read: EMA) you could run a meta-analysis of the two pivotals. At least one of them must have demonstrated BE. You can try.

“If for a particular formulation at a particular strength multiple studies have been performed some of which demonstrate bioequivalence and some of which do not, the body of evidence must be con­sidered as a whole. […] The existence of a study which demonstrates bioequivalence does not mean that those which do not can be ignored. The applicant should thoroughly discuss the results and justify the claim that bioequivalence has been demonstrated. Alternatively, when relevant, a combined analysis of all studies can be provided in addition to the individual study analyses. It is not acceptable to pool together studies which fail to demonstrate bioequivalence in the absence of a study that does.”


Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Louis
☆    

NC, USA,
2015-11-27 18:02
(3044 d 20:04 ago)

@ Helmut
Posting: # 15679
Views: 5,996
 

 More information, please

Thank you all for the quick responses. I will post more in depth reply soon - just finished Thanksgiving with family and friends and heading home soon. :-) .


Edit: Full quote removed. Please delete everything from the text of the original poster which is not necessary in understanding your answer; see also this post! [Helmut]
UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,636 registered users;
112 visitors (0 registered, 112 guests [including 5 identified bots]).
Forum time: 14:06 CET (Europe/Vienna)

With four parameters I can fit an elephant,
and with five I can make him wiggle his trunk.    John von Neumann

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5