Lucas
★    

Brazil,
2014-07-31 00:24
(3528 d 14:09 ago)

Posting: # 13320
Views: 11,563
 

 Rivastigmine Transdermal Patch [General Sta­tis­tics]

Hello dear colleagues.

We are planning a BE study of a transdermal patch of Rivastigmine. This pharmaceutical form requires additional performance analysis, according to the FDA draft guidance published in 2010. The additional studies are: “Adhesion analysis” and a “Skin Irritation and Sensitization study”.
The adhesion is measured in scores - 0 = ≥ 90% adhered; 1 = ≥ 75% to <90% adhered; 2 = ≥ 50% to < 75% adhered; 3 = > 0% to < 50%; and 4 = 0% adhered – at different time points (i.e. 6, 12 and 24 hours after attachment of the patch).
FDA requires a comparison between the scores of test and reference treatments (page 2): “The adhesion evaluation of the active test product and RLD must demonstrate that the upper bound of the one-sided 95% CI of the mean adhesion score for the test product minus 1.25 times the mean adhesion score for the RLD must be less than or equal to 0.” And also states (page 3): “The proportion of subjects with a meaningful degree of detachment should be no higher for the test product than for the RLD, and de­tach­ment should not occur earlier in the application period for the test than for the RLD.”
Let’s leave the Skin Irritation and Sensitization study aside for now. Any of u guys have experience with this kind of statistical analysis? Since the variable is a score, we do not know what would be a plausible test to be applied for the calculation of the CI. Also, how would one demonstrate what is asked in the phrase described above (of page 3)? Is it a non-superiority analysis? If so, how would that be?
The study is for submission in ANVISA, which does not have guidelines on that, and seems to be more willing to use FDA’s guidance as their reference.
ElMaestro
★★★

Denmark,
2014-07-31 01:57
(3528 d 12:36 ago)

@ Lucas
Posting: # 13321
Views: 9,771
 

 Rivastigmine Transdermal Patch

Hi Lucas,

confusing that they ask for means and not medians. I am nevertheless inclined to say you should use non-parametrics (Wilcoxon signed ranks / Hodges-Lehman) for requirement #1.

A potential way forward to requirement #2 is the construction of confidence intervals based on proportions of score 1 and above from binomial testing.
It worries me a little with several time points. As I see it this means several tests must be done which can be rather nasty. It makes power calculations quite difficult prior to planning the pivotal trial.

Those are guesses from my side. I have not been involved in Rivastigmine patch developments.
Best of luck.

Pass or fail!
ElMaestro
Lucas
★    

Brazil,
2014-07-31 16:00
(3527 d 22:33 ago)

(edited by Lucas on 2014-07-31 18:59)
@ ElMaestro
Posting: # 13325
Views: 9,828
 

 Rivastigmine Transdermal Patch

Hi ElMaestro!

Thanks for the reply.

❝ confusing that they ask for means and not medians.


Yes! Mean score does not make much sense, working with medians would be better. I found, yesterday, this Public Assessment Report from EMA where the Wilcoxon Signed Rank Test was used (Pages 13 and 14), but seems like no CI was assessed as FDA requests. I'm not really sure on how to construct a 95% CI for MSt-1.25*MSr (where MSt= Mean Score for test and MSr= Mean Score for reference).

❝ A potential way forward to requirement #2 is the construction of confidence intervals based on proportions of score 1 and above from binomial testing.


You mean classify the scores as meaningful (i.e. 2 or higher) and not meaningful (i.e. 1 or lower)?

❝ As I see it this means several tests must be done which can be rather nasty.


That worries me very much also.

I think that the Wilcoxon SR test, as it was used in the Actavis study linked above, would be a good way do compare T vs R, since it also evaluates (indirectly) the adhesion through time. What do you think?

Thanks in advance.

Lucas


UPDATE

I couldn't reproduce the p-value described in the Actavis' study (0.0882) in R, using wilcox.test(), can anybody help me with that?
ElMaestro
★★★

Denmark,
2014-07-31 21:22
(3527 d 17:11 ago)

@ Lucas
Posting: # 13328
Views: 9,689
 

 Rivastigmine Transdermal Patch

Hi Lucas,

❝ I found, yesterday, this Public Assessment Report from EMA where the Wilcoxon Signed Rank Test was used (Pages 13 and 14), but seems like no CI was assessed as FDA requests. I'm not really sure on how to construct a 95% CI for MSt-1.25*MSr (where MSt= Mean Score for test and MSr= Mean Score for reference).


The guideline can be read in two ways. Easiest and perhaps compliant is to construct a 95% CI for MSt and compare it with the border 1.25MSr.

❝ You mean classify the scores as meaningful (i.e. 2 or higher) and not meaningful (i.e. 1 or lower)?


I thought meaningful non-adhesion was 1 or higher.

❝ I think that the Wilcoxon SR test, as it was used in the Actavis study linked above, would be a good way do compare T vs R, since it also evaluates (indirectly) the adhesion through time. What do you think?


Sounds ok to me, if it is ok to regulators too is of course the million dollar question. Or perhaps we should speak of reals in this case ;-)

I just looked at the PAR you referred to: How on earth is table 11.4.8 to be interpreted? Can you help?

Pass or fail!
ElMaestro
Lucas
★    

Brazil,
2014-07-31 22:54
(3527 d 15:39 ago)

@ ElMaestro
Posting: # 13329
Views: 9,609
 

 Rivastigmine Transdermal Patch

Hello

❝ ...How on earth is table 11.4.8 to be interpreted?


The adhesion was evaluated at 9 time points during 24h for all subjects and the cummulative score was calculated, obtaining the CAS (Cummulative Adhesion Score) for each subject and treatment, which is the sum of all 9 scores.
We would have 9 evaluations for each period of each subject, each evaluation with a score from 0 to 4. Remember that the score will only increase from one time point to the next. These 9 scores would be cummulated to obtain the CAS. The values observed for CAS were 0 – 10 in this study. This table shows that, for instance, 22 of the 37 subjects had a CAS of 0 for the test treatment, meaning that the score for all the 9 time points of these subjects was 0.

But I just realised that I can't analyze the data without knowing the CAS for test and reference per subject. I was assuming that, for instance, the same subject who had a CAS of 10 for test had a CAS of 9 for reference, but I can't assume that. Very amateur mistake... :-(
lechia
☆    

C of U,
2014-10-07 21:26
(3459 d 17:07 ago)

@ Lucas
Posting: # 13667
Views: 9,126
 

 Rivastigmine Transdermal Patch

❝ ❝ ❝ ❝ And also states (page 3): “The proportion of subjects with a meaning­ful degree of detachment should be no higher for the test product than for the RLD, and detachment should not occur earlier in the application period for the test than for the RLD.”


You can do a dichotomized analysis for the difference in proportion of subjects with a meaningful degree of detachment vs. not for test versus reference at the last time point. Then, if 90% CI contains 0, you can assume non-inferiority.
Lucas
★    

Brazil,
2015-02-23 18:53
(3320 d 18:40 ago)

@ Lucas
Posting: # 14493
Views: 8,543
 

 Rivastigmine Transdermal Patch

I would like to re-open this thread, in light of the final version of EMA's guidance on Transdermal Patches (link here). This document guide us to the "Guideline on the Pharmacokinetic and clinical evaluation of modified-release dosage forms" to describe with details how to conduct Skin Irritation and Sensitization study and Adhesion study(annex I and annex II)
Let us talk about the Adhesion bit. EMA states: "The adhesion should be measured as the percentage of area that remains adhered at the end of the dosing interval." (my emphasis)
By that I assume that the variable of adhesion will be continuous. So I ask: how would one measure that with precision? In the draft of this guidance they suggested to categorize and treat as scores (similar to FDA's draft guidance) but with a statistical analysis that didn't make sense (now it makes more sense).
I've seen this method but I'm not sure if it is applicable, since it was suggested to use for scores (categorical data) and not for continuous. The precision in that method of assessment of adhesion would enough to treat the variable as continuous?
Anyone know a better method of assessment?
UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,639 registered users;
100 visitors (0 registered, 100 guests [including 7 identified bots]).
Forum time: 13:34 CET (Europe/Vienna)

Nothing shows a lack of mathematical education more
than an overly precise calculation.    Carl Friedrich Gauß

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