kumarnaidu
★    

Mumbai, India,
2014-11-10 09:16
(3426 d 00:18 ago)

Posting: # 13848
Views: 14,177
 

 Draft FDA guidance on Methylphenidate Hydrochloride [Regulatives / Guidelines]

Dear All,

Draft FDA guidance on Methylphenidate Hydrochloride Released: Draft Guidance on Methylphenidate Hydrochloride

Kumar Naidu
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-11-10 14:57
(3425 d 18:38 ago)

@ kumarnaidu
Posting: # 13850
Views: 13,078
 

 Five PK metrics + S×F interaction

Hi Kumar,

THX for posting! These requirements will be tough to meet.
  • Sep 2012
    • 2×2 cross-over studies
    • Fasting
      Cmax, AUC0-T, AUCT-t, AUC; where T = 3 h.
    • Fed
      As above, but T = 4 h.
  • Nov 2014
    • 4-period fully replicated studies
    • Fasting
      Cmax, AUC0-T1, AUCT1-T2, AUCT2-T3, AUC; where T1 = 3 h, T2 = 7 h, and T3 = 12 h.
    • Fed
      As above; but T1 = 4 h and T2 = 8 h.
    • In both studies estimate the subject-by-formulation interaction σ²D. Its upper 95% CL should be ≤0.03.
Submit your comments to regulations.gov until Jan 5, 2015.

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
AngusMcLean
★★  

USA,
2014-11-11 20:26
(3424 d 13:09 ago)

@ Helmut
Posting: # 13860
Views: 12,910
 

 Five PK metrics + S×F interaction

Helmut,

can the (subject by formulation) interaction be done in Phoenix WinNonlin? I wonder if it is in the Pharsight templates, which are being redeveloped for the new edition of Phoenix (6.4)?

Angus
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-11-11 22:05
(3424 d 11:30 ago)

@ AngusMcLean
Posting: # 13862
Views: 12,920
 

 Phoenix

Hi Angus,

❝ Can the (subject by formulation) interaction be done in Phoenix WinNonlin?


You don’t need the templates. Run conventional ABE with the default setup:

Fixed Effects: Sequence+Period+Formulation
Variance Structure > Random 1:
Random Effects: Formulation
Variance Blocking Variables: Subject
Type: Banded No-Diagonal Factor Analytic(f), Number of factors = 2
Variance Structure > Repeated:
Repeated Specification: Period
Variance Blocking Variables: Subject
Group: Formulation
Type: Variance Components


Final Variance Parameters > lambda(2,2)_11 is the estimated subject×formulation interaction, Var(Period*Formulation*Subject)_21 is σ²wR, and Var(Period*Formulation*Subject)_22 is σ²wT. It should be possible to merge the degrees of freedom, get cinv(0.05, df), and calculate the upper bound directly. Have to check. If not, all the necessary data to follow Step 4 of the guidance are avail­able.

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
AngusMcLean
★★  

USA,
2014-11-12 20:25
(3423 d 13:10 ago)

@ Helmut
Posting: # 13863
Views: 12,816
 

 Phoenix

Thank you Helmut that appears to be fine:

The subject-by-formulation interaction variance is the variation in the average T and R difference in PK parameter among individuals.

The estimated lambda(2,2)_11 is the estimated variance for the subject×formulation interaction and is 8.3391924E-07 from the ABE Object in the Pharsight RSABE FDA template.

It seems we should be able to calculate the upper 95% confidence interval from the information we have. Now fundamentally exactly how do we do that? I am told by Ana that this was never included in the templates.

Angus
AngusMcLean
★★  

USA,
2014-11-21 01:28
(3415 d 08:06 ago)

@ Helmut
Posting: # 13896
Views: 11,997
 

 Phoenix

Helmut: Again I looked at he subject by formulation interaction from the Pharsight progesterone template it is a rather small number {8.3391924E-07 from the ABE Object in the RSABE FDA template, final variance paramaters tab}

This number is taken from "Final Variance Parameters > lambda(2,2)_11 is the estimated subject×formulation"

I am wondering am I misunderstanding the outpput. Surely it is way too low, I looked at a stick plot of the repsonse variable (AUC)


[image]

Please can you comment,

Angus
jag009
★★★

NJ,
2014-11-16 04:39
(3420 d 04:56 ago)

@ Helmut
Posting: # 13870
Views: 12,577
 

 Five PK metrics + S×F interaction

Hi Helmut,

❝ THX for posting! These requirements will be tough to meet.


Laszlo E is not amused :-)

John
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-11-17 13:49
(3418 d 19:46 ago)

@ jag009
Posting: # 13871
Views: 12,479
 

 Acceptance limit for the upper 95% con­fi­dence bound?

Hi John,

❝ Laszlo E is not amused :-)


Yes. He also pointed to the last paragraph on page 6. On has to calculate the upper 95% con­fi­dence bound for σ²D – but the guidance doesn’t give an acceptance limit for it. Like in the 2001 stats guidance we only find “the recommended allowance for σ²D is 0.03”. Why struggle with its CL if only the estimate has to be assessed?

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
jag009
★★★

NJ,
2014-11-17 18:29
(3418 d 15:06 ago)

@ Helmut
Posting: # 13872
Views: 12,356
 

 Acceptance limit for the upper 95% con­fi­dence bound?

Hi Helmut,

Did he tell you that the calculation is not even correct or something? I don't remember his exact words but he said he proved something in his previous paper.
I will find..

John
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-11-18 04:56
(3418 d 04:39 ago)

@ jag009
Posting: # 13873
Views: 12,407
 

 Quoting László

Hi John,

❝ Did he tell you that the calculation is not even correct or something?


Yep: “[…] we demonstrated in our papers that the estimated interaction is related, and proportional, to the within-subject variance. Therefore, it does not make sense to set a critical value for its estimate or its confidence limit.”

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
AngusMcLean
★★  

USA,
2014-11-19 14:28
(3416 d 19:07 ago)

@ Helmut
Posting: # 13878
Views: 12,200
 

 Acceptance limit for the upper 95% con­fi­dence bound?

Helmut,

presumably they have decided not to set a limit, since right now they do not have the data to do it. Once they have submissions coming in from sponsors then they could look at the values then set a limit or alternatively drop the requirement.
kumarnaidu
★    

Mumbai, India,
2014-12-09 09:36
(3396 d 23:58 ago)

@ Helmut
Posting: # 14032
Views: 11,121
 

 Five PK metrics + S×F interaction

Hello all,
One thing I want to confirm here that if for any of the parameter Swr >0.294 observed then can we use scaled approach there?:confused:

Kumar Naidu
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-12-09 14:18
(3396 d 19:16 ago)

@ kumarnaidu
Posting: # 14033
Views: 11,089
 

 No RSABE

Hi Kumar,

❝ One thing I want to confirm here that if for any of the parameter Swr >0.294 observed then can we use scaled approach there?:confused:


I don’t think so. Quoting the guidance:

The 90% confidence intervals of the geometric mean test/reference (T/R) ratios for the above five Cmax and AUC metrics (Cmax, AUC0-T1, AUCT1-T2, AUCT2-T3, AUC0-∞) should fall within the limits of 80–125%.

Not a single word about reference-scaling. Given the intrasubject variabilities of these metrics I don’t see a justification.

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
kumarnaidu
★    

Mumbai, India,
2014-12-09 15:19
(3396 d 18:16 ago)

@ Helmut
Posting: # 14034
Views: 11,142
 

 No RSABE

Hi Helmut,
Actually one of my friend recently did this study and according to him he is getting ISCV around 32% for one of the partial AUCs and the upper limit of T/R ratio for that parameter is 127%. And he planning to use scaled approach.

One more thing for one of the parameter he is getting σ²D >0.03 and the T/R ratio within 80 to 125 (borderline pass). How to justify here ? :-|
(I dont know how he completed this study in one month)

Kumar Naidu
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-12-09 15:50
(3396 d 17:45 ago)

@ kumarnaidu
Posting: # 14035
Views: 11,330
 

 No RSABE

Hi Kumar,

❝ Actually one of my friend recently did this study and according to him he is getting ISCV around 32% for one of the partial AUCs and the upper limit of T/R ratio for that parameter is 127%.


Tell your friend that he should evaluate the study according to the protocol. Since the study is already completed it was performed according to the ”old” guidance, right? Not a word about RSABE. Upper CL 127% = study failed to show BE. Interesting that he performed a replicate design.

❝ And he planning to use scaled approach.


In an attempt to “safe” a study which failed according to the protocol? Maybe. See the RTR guidance, Section “Failed In Vivo BE Studies”:

FDA will refuse-to-receive an ANDA if only a failed in vivo BE study is submitted. FDA regu­la­tions require applicants to submit information on failed BE studies. Typi­cally, a failed study is one that does not satisfy the 90% con­fi­dence interval (CI) criterion (i.e., falls outside of the 0.8–1.25 acceptance criterion limits) for either area under curve (AUC) or the peak plasma concentration (Cmax) parameter. If this occurs for highly variable drug products, the applicant should submit a study using a replicate study design and analyze data using a reference-scaled average (RSA) approach for the failed parameter. Alternatively, applicants should consult the BE recom­men­da­tions webpage for product-specific study information or contact OGD’s Division of Bioequivalence via a BE Guidance Request for further guidance if needed.


Some thoughts:
  • The study failed according to the ­“old” product-specific guidance.
  • The study was already performed in a replicate design which would allow RSABE.
  • From the wording I get the impression that the FDA wants to see another study. Duno whether FDA would allow a post hoc switch from ABE to RSABE in the same study.
  • But: The current guidance does not mention RSABE as well.
  • I would contact the OGD via controlled correspondence.

❝ One more thing for one of the parameter he is getting σ²D >0.03 and the T/R ratio within 80 to 125 (borderline pass). How to justify here ? :-|


What do you mean by “justify”? That’s really bad luck.

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
drgunasakaran1
★★  
avatar

2014-11-10 19:29
(3425 d 14:06 ago)

@ kumarnaidu
Posting: # 13852
Views: 12,915
 

 Five PK metrics + S×F interaction

Dear Kumar,
Thanks for sharing the revised guidance.
As Mr Helmut rightly said, they are asking for additional partial AUCs which will be difficult to match and suggested to estimate subject-by-formulation interaction..!
Will this subject-by-formulation estimation applies for other Bimodal release formulations too.?:crying:

Dr S Gunasakaran MBBS MD
Disclaimer: The replies/posts are my personal opinions and it does not represent my company views on the same.
jag009
★★★

NJ,
2014-11-11 19:04
(3424 d 14:30 ago)

@ drgunasakaran1
Posting: # 13858
Views: 12,810
 

 Five PK metrics + S×F interaction

I don't like the subject x formulation interaction evaluation. What if a sponsors conducted the studies based on last guidance (2way studies), there is no way to perform the sub x form interaction analysis (?) even if they managed to pass every PK parameters...

Why do we need to evaulate the sub x form interaction??? Looks to me like an overkill.

John
AngusMcLean
★★  

USA,
2014-11-11 20:22
(3424 d 13:13 ago)

@ drgunasakaran1
Posting: # 13859
Views: 12,882
 

 Five PK metrics + S×F interaction

I think it may be applied to other MP bimodal formulations. I would study the fundamental release aspects of Concerta in vitro and invivo. Then I would think of analogous metrics as is appropriate for your drug of interest. In other words make a plan in case you are asked to include partial exposure metrics.

If it applies to Concerta I myself can see no reason why this would not be applied to other bimodal drugs.

Any comments,

Angus
jag009
★★★

NJ,
2014-11-11 21:08
(3424 d 12:26 ago)

@ AngusMcLean
Posting: # 13861
Views: 12,845
 

 Five PK metrics + S×F inter­action

I kind of understand the logic for this with Concerta due to its zero order release ER component. Whereas for other bimodals like Adderall XR, Metadate CD, it's different.

John
jag009
★★★

NJ,
2014-11-13 21:17
(3422 d 12:18 ago)

@ jag009
Posting: # 13864
Views: 12,814
 

 Tada!

The biggie starts!

http://finance.yahoo.com/news/mallinckrodt-plc-responds-fda-expected-110000949.html

Wonder if we will see Adderall XR generic on the same topic soon.
AngusMcLean
★★  

USA,
2014-11-14 00:06
(3422 d 09:28 ago)

@ jag009
Posting: # 13866
Views: 12,621
 

 Tada!

I am thinking "yes". But the metrics should be appropriate for XR (double- pulsed delivery), since XR comprises an IR bead plus an EC bead releasing at pH 5.5.
nobody
nothing

2014-11-19 10:30
(3416 d 23:05 ago)

@ AngusMcLean
Posting: # 13875
Views: 12,452
 

 Tada!

Clix me

What I think is interesting is that the FDA started BE AND clin. study, apparently not fully trusting the concept of BE anymore regarding some special drugs. Comes down to the discussion we had lately on how to set up "appropriate" BE limits in general (and how products tested according to these limits perform under real-life conditions)...

Kindest regards, nobody
AngusMcLean
★★  

USA,
2014-11-19 14:42
(3416 d 18:52 ago)

@ nobody
Posting: # 13879
Views: 12,247
 

 Tada!

What they are doing with methylphenidate (Concerta in this case) is linking the established onset and duration of the drug effect (12 hours) with corresponding BE AUC intervals across the 12 hour period. As far as I know they generally stick with the usual 80-125 intervals. Perhaps in Europe, in general, there is more flexibility on occasions for the intervals?


Angus
nobody
nothing

2014-11-19 14:55
(3416 d 18:39 ago)

@ AngusMcLean
Posting: # 13880
Views: 12,585
 

 Tada!

Other range of 90%-111% only for narrow therapeutic index drugs in Europe

The truncated AUC concept is clear, but they don't trust a BE trial in healthy adults at the FDA and started a PK-PD study in children, that's the point where I say: The concept of BE is not accepted any more in general.

Kindest regards, nobody
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-11-19 16:44
(3416 d 16:51 ago)

@ nobody
Posting: # 13884
Views: 12,347
 

 ABEL

Hi nobody & Angus,

❝ Other range of 90%-111% only for narrow therapeutic index drugs in Europe


AFAIK a generic of Concerta passed a centralised procedure a while ago. EMA accepted ABEL (i.e., widening of the acceptance range) not only for Cmax, but also for the (first only?) pAUC. The same is mentioned in the MR-Draft (line 810). Let’s wait for the final version.

I’m pretty sure that most (all?) other MR formulations in the EU were hybrid appli­ca­tions: BA-studies against IR MPH given twice (τ 4 hours) plus one (or more) therapeutic equivalence studies in children. BTW, approved MR for­mu­la­tions are not inter­change­able.1,2


    References:
  1. Haessler F, Tracik F, Dietrich H, Stammer H, Klatt J. A pharmacokinetic study of two modified-release methylphenidate formulations under different food conditions in healthy volunteers. Int J Clin Pharmacol Ther. 2008;46(9):466–76. PMID 18793577
  2. Schütz H, Fischer R, Großmann M, Mazur D, Leis HJ, Ammer R. Lack of bioequivalence between two methylphenidate extended modified release formulations in healthy volunteers. Int J Clin Pharmacol Ther. 2009;47(12):761–9. doi 10.5414/CPP47761
PS @nobody: Any particular reason why you linked the draft BE-GL?

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
nobody
nothing

2014-11-19 17:05
(3416 d 16:30 ago)

@ Helmut
Posting: # 13886
Views: 12,174
 

 ABEL

Why not interchangeable when MA was granted based on BE-studies for at least one formulation? Could you provide some details on the reasons? Thanx nobody

PS: Found somefink interesting, a strange company called BEBAC has apparently been digging deep into methylphenidate in the past...

http://www.ema.europa.eu/docs/en_GB/document_library/Referrals_document/Methylphenidate_Hexal/WC500156885.pdf

Kindest regards, nobody
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-11-19 17:59
(3416 d 15:36 ago)

@ nobody
Posting: # 13887
Views: 12,326
 

 ABEL

Hi nobody,

❝ Why not interchangeable when MA was granted based on BE-studies for at least one formulation? Could you provide some details on the reasons?


This particular one was BE to Concerta and no clinical studies were performed. The other products stand on their own (clinical data) and these two papers show lack of BE between them.

❝ PS: Found somefink interesting, a strange company called BEBAC has apparently been digging deep into methylphenidate in the past...


Well yes, though I didn’t support Hexal in this project. ;-) My presentations are licensed [image] and they quoted the source correctly. But they used it for com­mer­cial pur­poses (violating the license). Shall I sue them?

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
nobody
nothing

2014-11-19 18:10
(3416 d 15:25 ago)

@ Helmut
Posting: # 13888
Views: 12,104
 

 ABEL

They didn' ask or pay for that? WOW, that's a bold move... at the best. But pharma-world is small in Europe, so... YOUR choice ;-) Maybe call Ms T-F?

Kindest regards, nobody
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-11-19 18:16
(3416 d 15:19 ago)

@ nobody
Posting: # 13889
Views: 12,172
 

 ABEL

Hi nobody,

❝ They didn' ask or pay for that?


Neither nor.

❝ WOW, that's a bold move... at the best.


Yep. :angry:

❝ But pharma-world is small in Europe, so... YOUR choice ;-) Maybe call Ms T-F?


Good idea!

PS: A cut-off time of two hours in fed state is crap.

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
AngusMcLean
★★  

USA,
2014-11-19 18:45
(3416 d 14:49 ago)

@ Helmut
Posting: # 13890
Views: 12,152
 

 ABEL

I think the fed within subject variability chart is interesting: it shows the variability at the earlier time intervals in the fed state is huge and that makes the point very well. I am thinking that the food delays gastric emptying so the formulations are retained in the stomach ~ 3 hours prior to transiting to the small intestine. {would have thought that Concerta would be OK, since it has an IR overcoat component}.

I wonder how the within subject variability in the fasted state compares with the fed state for such formulations at the corresponding time intervals.

Angus
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-11-19 21:15
(3416 d 12:20 ago)

@ AngusMcLean
Posting: # 13891
Views: 12,201
 

 Difficult formulations

Hi Angus,

❝ […] it shows the variability at the earlier time intervals in the fed state is huge and that makes the point very well. I am thinking that the food delays gastric emptying so the formulations are retained in the stomach ~ 3 hours prior to transiting to the small intestine.


Exactly. That’s actually the idea behind formulations with differently coated pellets.

❝ {would have thought that Concerta would be OK, since it has an IR overcoat component}.


Concerta shows a smaller food effect, but we see some shift in osmotic pumps as well.

❝ I wonder how the within subject variability in the fasted state compares with the fed state for such formulations at the corresponding time intervals.


That’s a difficult question. There are formulations with different release mecha­nisms on the market. All show a food effect (practically only on rate metrics), some don’t work in fasted state very well (dose dumping), most fail in vitro alcohol-induced dose dumping, variability of the early phase in fed state might be higher (most formulations) or lower (some drugs)… :blahblah:

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
nobody
nothing

2014-11-19 22:04
(3416 d 11:31 ago)

@ Helmut
Posting: # 13892
Views: 12,197
 

 Difficult formulations

…if you want to reduce food effect, multiple-unit formulations is usually a good try, but no guarantee.. ;-)

Kindest regards, nobody
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-11-19 22:49
(3416 d 10:46 ago)

@ nobody
Posting: # 13893
Views: 12,234
 

 Difficult formulations

Hi nobody,

❝ …if you want to reduce food effect, multiple-unit formulations is usually a good try,


… which AFAIK all non-osmotic pump formulations of MPH are.

❝ … but no guarantee.. ;-)


Xactly.

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
UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,640 registered users;
50 visitors (0 registered, 50 guests [including 2 identified bots]).
Forum time: 09:35 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