balakotu
★    

India,
2015-07-02 11:05
(3192 d 06:47 ago)

Posting: # 15013
Views: 9,160
 

 HC Proposed Policy on Bioequivalence Standards for HVDS [Regulatives / Guidelines]

d_labes
★★★

Berlin, Germany,
2015-07-02 11:19
(3192 d 06:33 ago)

@ balakotu
Posting: # 15015
Views: 8,581
 

 HC Proposed Policy on Bioequivalence Standards for HVDS

Dear Kotu,

THX for that link.

Interesting! EMA approach but allowed for AUC. Since Cmax doesn't need a CI criterion at all, only a point estimate restriction is imposed.

Regards,

Detlew
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-02 15:20
(3192 d 02:32 ago)

@ balakotu
Posting: # 15017
Views: 8,380
 

 Bloody partial replicate again

Hi Kotu,

THX for the information.

Why the heck:

The test product (T) should be administered either once in a 3-period design (RTR, TRR, RRT) or twice in a 4-period design (TRTR, RTRT).

No full replicate 3-period design (TRT|RTR)? See this lengthy thread. Disappointing.

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
Lucas
★    

Brazil,
2015-07-02 15:35
(3192 d 02:18 ago)

@ Helmut
Posting: # 15019
Views: 8,391
 

 Bloody partial replicate again

Hello my friends

❝ Why the heck:[...]



Yes! I do not like partial replicate designs also. I have a feeling that, taking those converging problems aside, having 2x more data of one treatment than the other is not good for the estimations. Seems like you give more chances for one treatment to show its "real PK profile". Could I say that?

Thx for the link.

Lucas
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-02 18:41
(3191 d 23:12 ago)

@ Lucas
Posting: # 15022
Views: 8,417
 

 Which model?

Hi Lucas,

❝ I do not like partial replicate designs also. I have a feeling that, taking those converging problems aside, …


To me it is not clear which statistical model HC prefers (see this post). Sounds like a mixed-effects model (FDA’s 2001 guidance, FDA’s progesterone guidance / unscaled ABE, EMA’s Method C) to me. Only then one may run into convergence issues. No problem with EMA’s crippled models (Methods A [subjects fixed] and B [subjects random]).

❝ … having 2x more data of one treatment than the other is not good for the estimations. Seems like you give more chances for one treatment to show its "real PK profile". Could I say that?


You can, of course. This forum is not censored. :-D

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
Lucas
★    

Brazil,
2015-07-02 20:06
(3191 d 21:47 ago)

@ Helmut
Posting: # 15025
Views: 8,335
 

 Which model?

Hi Helmut!

❝ You can, of course. This forum is not censored. :-D


Maybe ESL is to blame here. :-D

What I meant is "does that statement makes sense?".

Thx.
Lucas Teixeira
drgunasakaran1
★★  
avatar

2015-07-02 19:08
(3191 d 22:44 ago)

@ balakotu
Posting: # 15023
Views: 8,391
 

 HC Proposed Policy on Bioequivalence Standards for HVDS

Dear Mr Kotu,

Thanks for sharing the link.
It's surprising to see that the HVDP classification is based on the within subject coefficient of variation of AUC alone and not based on AUC and/or Cmax.

Dr S Gunasakaran MBBS MD
Disclaimer: The replies/posts are my personal opinions and it does not represent my company views on the same.
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-02 19:37
(3191 d 22:16 ago)

@ drgunasakaran1
Posting: # 15024
Views: 8,311
 

 AUC and Cmax

Dear Dr. S. Gunasakaran,

❝ It's surprising to see that the HVDP classification is based on the within subject coefficient of variation of AUC alone and not based on AUC and/or Cmax.


Last year HC considered FDA’s approach (see here). On many previous occasions HC stated that they don’t see any need for changing the requirement for Cmax since for more than two decades only the T/R-ratio had to be contained within 80–125%, which – in HC’s opinion – was never an obstacle con­sidering the necessary sample sizes. László Endrényi and László Tóthfalusi showed many times that for high CVs (say ≥50%) only the restriction on the point estimate leads the BE-decision (irrespective of the scaling method).

Therefore, I understand the logic behind:
  • Keep the current method for Cmax – which served well for ages.
  • Allow reference-scaling of AUC.

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
d_labes
★★★

Berlin, Germany,
2015-07-03 10:21
(3191 d 07:32 ago)

@ balakotu
Posting: # 15028
Views: 8,174
 

 HC for HVDs

Dear all,

another interesting detail: All (widened) acceptance ranges given with one decimal, even the conventional 80 - 125%.

Regards,

Detlew
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-03 13:55
(3191 d 03:58 ago)

@ d_labes
Posting: # 15032
Views: 8,211
 

 HC for HVDs

Dear Detlew,

that’s consistent with HC’s other current requirements (generally 80.0–125.0%, critical dose drugs 90.0–112.0%).

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
d_labes
★★★

Berlin, Germany,
2015-07-03 14:53
(3191 d 03:00 ago)

@ Helmut
Posting: # 15033
Views: 8,123
 

 Rounding

Dear Helmut,

thanks for reminding me.

❝ critical dose drugs 90.0–112.0%).


Although this is a bit of generous rounding to one decimal :cool:

Regards,

Detlew
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-03 16:49
(3191 d 01:03 ago)

@ d_labes
Posting: # 15035
Views: 8,188
 

 Acceptable difference (BE history)

Dear Detlew,

❝ ❝ critical dose drugs 90.0–112.0%).


❝ Although this is a bit of generous rounding to one decimal :cool:


I told Eric Ormsby at the last BioInternational conference (London, October 2008) that on the long run the mean of generic NTIDs will be \(\sqrt{0.900 \times 1.120} = 1.003992 \ldots\). IIRC he replied “We can stand that. 112% is just easier to remember.”
When discussions were hot in the mid-1980s (switching from untransformed to log-trans­formed data) we had a width of the acceptance range of 0.40 (80–120%) based on an acceptable Δ of 0.20:
  • Some people proposed θ1 = 1–Δ (0.80) and θ2 = 1/θ1 (1.25) or ±ln(0.80).
  • Others argued that with this method the AR in the original scale (0.45) will be wider than the previous standard (0.40). They suggested an even more conservative approach:
    θ2 = 1+Δ (1.20) and θ1 = 1/θ1 (0.83) or ±ln(0.83).
  • A minority suggested to modify Δ in such a way (to 0.1802…) that the AR keeps the previous width, i.e., to θ1 0.8198 and θ2 1.2198.
We all know which party won – leading to limits “which are easy to remember”. Yes, this was at the end the unbeatable argument.:-D

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
ElMaestro
★★★

Denmark,
2015-07-03 15:02
(3191 d 02:50 ago)

@ Helmut
Posting: # 15034
Views: 8,106
 

 Off topic

Hi all,

I am still baffled by scaling, whichever way it is done. Granted, I do not understand much of BE and in particular I am struggling with scaling. I am baffled for this reason:

As an alternative to scaling we might in stead relax the alpha and keep accepting and rejecting the same products. We would achieve not just roughly the same but exactly the same because scaling keeps the GMR unaffected. A practical matter is of course to identify the mathematical relationship between the scaling constant of acceptance range and scaling of alpha but this is nothing more than a practicality (find a closed form or chose empery, whatever works for you).

I think the only reason not to formulate the problem (=acceptance vs rejection) in terms of alpha relaxation but rather in terms of widening of CI limits is that if we formulate it as a widening of alpha level then we need to admit explicitly what scaling involves: An occasional increase in type I error.

Prove me wrong, please.

Pass or fail!
ElMaestro
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-03 17:38
(3191 d 00:15 ago)

@ ElMaestro
Posting: # 15036
Views: 8,244
 

 Playing around with α (not exactly brilliant)

Ahoy!

❝ As an alternative to scaling we might in stead relax the alpha and keep accepting and rejecting the same products…


I disagree. The BE-problem is originally stated in terms of the maximum acceptable difference (Δ), where Δ is considered to be clinically not relevant. This Δ leads to the acceptance range (see the history lesson above).

If a a drug/drug product is highly variable it will be very difficult to demonstrate BE of the reference to itself given the con­ven­tio­nal Δ of 20%. On the other hand, practically all HVDs/HVDPs have a flat dose response curve (i.e., even a large Δ is not expected to be clinically relevant). Hence, it makes sense to widen the accep­tance range (either to a fixed value in some jurisdictions or scaled based on the swR in others).
If we modify the AR, we can still keep the patient’s risk. I know of only one regulation where the α is mo­di­fied. ANVISA requires for NTIDs the conventional AR but assessed by a 95% CI. IMHO, that’s a flawed approach. Yes, the risk will be lower. But which risk? To have 2.5% of the patient-population with a BA of <80% and 2.5% with >125%. Can be nasty for NTIDs. I don’t get the “idea”.

❝ I think the only reason not to formulate the problem (=acceptance vs rejection) in terms of alpha relaxation but rather in terms of widening of CI limits is that if we formulate it as a widening of alpha level then we need to admit explicitly what scaling involves: An occasional increase in type I error.


An inflated TIE could easily (!) be avoided by either pre-specifying a lower α (0.025 or 0.0304 depen­dent on the design) or adjusting α based on swR. Time to publish sumfink.
Even then, one big problem – common to all RSABE-methods – remains open. Products are approved according to different standards. Furthermore, the procedure is not reversible. In ABE we can recode T with R and the decision (pass/fail) will always be the same. This is not the case in RSABE if swR ≠ swT (of course we need a fully replicated design to show that).

I think that the two Lászlós in one of their papers mentioned that it would be desirable that agencies pool swR-data and publish a fixed widened acceptance range in product-specific guidelines. In such a case the Null-Hypothesis is no more modified “in face of the data” and no inflation of the TIE is pos­si­ble. No replicate designs, no fiddling around with crappy models. A conventional 2×2 would do the job (back to the roots).
This is what I hope for (ha-ha): RSABE as a kind of “transition state” until enough data are collected allowing regulators to come up with justified recommendations.
Will never happen. I know, I know.

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
ElMaestro
★★★

Denmark,
2015-07-03 18:35
(3190 d 23:18 ago)

@ Helmut
Posting: # 15037
Views: 8,004
 

 Playing around with α (not exactly brilliant)

Hi Helmut,

I see your points, however, but how did they relate to my point? I reckon we address two different aspects. It is likely a question of bad explanations on my part.
Consider this:
  • For any CV measured you can define a revised acceptance range while keeping the originally applied alpha (5%).
  • For any CV measured I can define a revised alpha that keeps the original acceptance range in place (80.00%-125.00%), and which achieves exactly the same (it rejects and accepts exactly the same products as your approach).
I can probably work out an example if you need one? Or someone well versed with stats can. Think of the principle, rather than how to exactly describe the relationship between scaled acceptance range and applied alpha while keeping the rejections/acceptances same.

My point was/is that changing the acceptance range for the CI in mathematical terms does exactly the same as keeping the acceptance range while changing the applied alpha.

Pass or fail!
ElMaestro
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-03 18:44
(3190 d 23:08 ago)

@ ElMaestro
Posting: # 15038
Views: 8,025
 

 Playing around with α (not exactly brilliant)

Hi ElMaestro,

❝ – For any CV measured I can define a revised alpha that keeps the original acceptance range in place (80.00%-125.00%), and which achieves exactly the same (it rejects and accepts exactly the same products as your approach).


Yep, but (big but): Your α will necessarily be (much) larger than 0.05. Let’s not warm-up the dispute RA Fisher had with Jerzy Neyman. At least I don’t want to go there. ;-)

❝ I can probably work out an example if you need one?


THX, I do understand what you mean.

❝ My point was/is that changing the acceptance range for the CI in mathematical terms does exactly the same as keeping the acceptance range while changing the applied alpha.


Sure. Would you prefer to say “The risk for the patient to have a BA outside 80–125% of the reference is 31.42%”?

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
d_labes
★★★

Berlin, Germany,
2015-07-03 23:11
(3190 d 18:42 ago)

@ Helmut
Posting: # 15039
Views: 7,975
 

 OT – Philosophers on duty

Gents (Helmut & Ol'Pirate)!

Relax and enjoy the mediteran weather in Europe.
Seems we get Greek conditions here, hopefully not the monetary one :cool:.

Whatch the scientific quote for this day (I will copy it for you since next day it is gone):

Statistics: The only science that enables different experts using the same figures to draw different conclusions.
Evan Esar


Full stop.

SCRN :-D.

Regards,

Detlew
UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,638 registered users;
75 visitors (0 registered, 75 guests [including 6 identified bots]).
Forum time: 16:53 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