Shatha
☆    

2022-04-04 13:35
(86 d 05:30 ago)

Posting: # 22906
Views: 897
 

 SAS: Hel­mert trans­for­ma­tion [Software]

Hello

I have 2 treatments, 2 periods, crossover, multiple dose, steady state study.
The last 3 pre-dose concentrations should be used for steady state verification for each subject in each period.
I need your help to perform this verification using Helmert transformation method via SAS software, any specific SAS code?

Thank you.


Edit: Category changed; see also this post #1[Helmut]
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2022-04-04 14:23
(86 d 04:42 ago)

@ Shatha
Posting: # 22907
Views: 784
 

 Hel­mert trans­for­ma­tion?

Hi Shatha,

» I have 2 treatments, 2 periods, crossover, multiple dose, steady state study.
» The last 3 pre-dose concentrations should be used for steady state verification for each subject in each period.

You can’t verify anything in science.
In statistics you can only try reject a properly stated [image] null hypothesis – which is in your case what? Any test at level \(\small{\alpha}\) has an unavoidable false positive rate. Do you want to exclude \(\small{(100\times\alpha)\%}\) of subjects although they are in (pseu­do) steady state?

» I need your help to perform this verification using Helmert transformation method …

Do you have a reference demonstrating its application in PK? Never came across this approach in 40+ years. See there why statistical tests are nonsense – essentially because we don’t have a proper null.

Dif-tor heh smusma 🖖 [image]
Helmut Schütz
[image]

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

2022-04-05 12:00
(85 d 07:04 ago)

@ Helmut
Posting: # 22910
Views: 802
 

 Hel­mert trans­for­ma­tion?

» You can’t verify anything in science.
» In statistics you can only try reject a properly stated [image] null hypothesis – which is in your case what? Any test at level \(\small{\alpha}\) has an unavoidable false positive rate. Do you want to exclude \(\small{(100\times\alpha)\%}\) of subjects although they are in (pseu­do) steady state?

I will not exclude any subjects from analysis, I just need to check achievement of steady state in compliance with EMA guideline. The guideline didn't specify the verification method.
I have reviewed the comments raised for this guidance, the proposed changes in page 53 and page 54 include visual inspection of the last three pre-dose concentrations to verify steady state achievement and the feasibility to use the last 2 pre-dose concentrations and the Ctau instead of the last 3 pre-dose concentrations.
EMA has evaluated the comments.
They mentioned in page 4 that clarifications on steady state evaluation was included in the final guidance following stakeholder comments. I searched for this final guidance but I didn't find it, the published effective version is still the one came into effect in 2015.
Is it feasible to verify steady state visually?

» » I need your help to perform this verification using Helmert transformation method
»
» Do you have a reference demonstrating its application in PK? Never came across this approach in 40+ years. See there why statistical tests are nonsense – essentially because we don’t have a proper null.

No, I don't have references. It was suggested as the best method for steady state verification by a reviewer during protocol preparation two years ago and accordingly documented in the protocol.

Thanks for your cooperation.
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2022-04-05 14:29
(85 d 04:36 ago)

@ Shatha
Posting: # 22911
Views: 730
 

 Steady-state achieve­ment

Hi Shata,

» » Any test at level \(\small{\alpha}\) has an unavoidable false positive rate. Do you want to exclude \(\small{(100\times\alpha)\%}\) of subjects although they are in (pseu­do) steady state?
»
» I will not exclude any subjects from analysis, I just need to check achievement of steady state in compliance with EMA guideline.

Under the premise of linear PK, the superposition principle holds, i.e., \(\small{AUC_{0-\infty}=AUC_{0-\tau}}\). In simple terms, we must compare bioavailabilites either after a single dose or in steady-state. If steady-state is not achieved, a comparison is not correct.
What will you do if your ‘check’ fails? Since you don’t want to exclude subjects, would you throw the entire study into the waste container?

» The guideline didn't specify the verification method.

Correct.
  • 5.1. Pharmacokinetic studies (p. 10)
    Whenever multiple dose studies are performed it should be demonstrated that steady state has been reached. Achievement of steady-state is assessed by comparing at least three pre-dose concentrations for each formulation, unless otherwise justified.
  • 6.1.1.2. Multiple dose studies (p. 19)
    Whether the steady-state has been achieved is assessed by comparing at least three pre-dose con­cen­tra­tions for each formulation. The apparent half-life should also be taken into account.
Section 5 is applicable for hybrid applications and Section 6 for generics. Don’t know why the half-life entered the scene here and not in Section 5 as well. If we cross the border of flip-flop PK (\(\small{k_\textrm{a}\leq k_\textrm{e}}\)), it will just take longer to reach steady state (i.e., designing the study based on \(\small{k_\textrm{e}}\) would be wrong).

» I have reviewed the comments raised for this guidance, the proposed changes in page 53 and page 54 include visual inspection of the last three pre-dose concentrations to verify steady state achievement and the feasibility to use the last 2 pre-dose concentrations and the Ctau instead of the last 3 pre-dose concentrations.
» EMA has evaluated the comments.

Yep. Note that though the overview of comments are dated with 17 October 2019 and according to the EMA’s website were published on 29 October 2019, the PDF was modified on 18 May 2020…
On 30 April 20141 Jan Neu­hau­ser of the Austrian agency AGES and mem­ber of the CHMP Pharmacokinetics Working Party announced that

»the overview of comments will be published within the next two weeks«.

5½ years two weeks. :thumb down:

» They mentioned in page 4 that clarifications on steady state evaluation was included in the final guidance following stakeholder comments.

If fail to find that in page 4 and anywhere else.

» I searched for this final guidance but I didn't find it, the published effective version is still the one came into effect in 2015.

That’s the final one.

» Is it feasible to verify steady state visually?

As I wrote in the article:

In 20132 members of the EMA’s Pharmacokinetics Working Party confirmed unanimously that a test is not expected by the agencies. ‘Demonstration of steady state’ should not be understood in the statistical sense. In­stead, individual pre-dose concentrations should be reported together with spaghetti plots and plots of geometric means. In the discussion it became clear that common sense should prevail.

In an earlier meeting3 nothing was mentioned at all.

For 10+ years I’m presenting only summary tables, individual and spaghetti plots. All regulatory assessors were fine with that.

» » » Helmert transformation method
» »
» » Do you have a reference demonstrating its application in PK?
»
» No, I don't have references. It was suggested as the best method for steady state verification by a reviewer during protocol preparation two years ago and accordingly documented in the protocol.

I can’t follow you here. You stated something in the protocol without knowing how to perform the method? Ask the reviewer for help and report back here. I’m curious.


  1. EMA/EGA. Joint workshop on the impact of the revised EMA Guideline on the Pharma­co­kinetic and Clinical Eva­lu­ation of Modified Release Dosage Forms. London, UK. 30 April 2014.
  2. EUFEPS. BABP Network Open Discussion Forum on the Revised European Guideline on Phar­ma­co­ki­netic and Clinical Evaluation of Mo­di­fied Release Dosage Forms. Bonn, Germany. 17–18 June 2013.
  3. EUFEPS. BABP Network Open Discussion Forum. Revision of BE Requirements for Modified Release Products Barcelona, Spain. 23–24 February 2011.

Dif-tor heh smusma 🖖 [image]
Helmut Schütz
[image]

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

2022-04-06 11:38
(84 d 07:26 ago)

@ Helmut
Posting: # 22912
Views: 678
 

 Steady-state achieve­ment

Hi Helmut

» What will you do if your ‘check’ fails? Since you don’t want to exclude subjects, would you throw the entire study into the waste container?

So, if the test isn't clearly requested, then I don't have to do it in order to avoid any consequences, right?

» As I wrote in the article:

In 20132 members of the EMA’s Pharmacokinetics Working Party confirmed unanimously that a test is not expected by the agencies. ‘Demonstration of steady state’ should not be understood in the statistical sense. In­stead, individual pre-dose concentrations should be reported together with spaghetti plots and plots of geometric means. In the discussion it became clear that common sense should prevail.

In an earlier meeting3 nothing was mentioned at all.
»
» For 10+ years I’m presenting only summary tables, individual and spaghetti plots. All regulatory assessors were fine with that.

I don't have access to the Open Forum Discussion, I tried to find reference no. 02 but I only found session 3 https://custom.cvent.com/6E33084F71EB4DD6B892656072736BFB/files/07323e6d219f47cdae68e89f1f508b6b.pdf, which isn't related to steady state achievement. Can you please provide the steady state related session to support protocol deviation/ amendment? Thank you

» I can’t follow you here. You stated something in the protocol without knowing how to perform the method? Ask the reviewer for help and report back here. I’m curious.

Please see the article and graph
[image]

Thank you
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2022-04-06 14:29
(84 d 04:35 ago)

@ Shatha
Posting: # 22913
Views: 647
 

 Steady-state achieve­ment

Hi Shatha,

» » What will you do if your ‘check’ fails? Since you don’t want to exclude subjects, would you throw the entire study into the waste container?
»
» So, if the test isn't clearly requested, then I don't have to do it in order to avoid any consequences, right?

Any confirmatory test has to have consequences. If a test is only exploratory, no problems. However, since it is a pre­re­qui­site for a valid comparison that (pseudo-)steady-state is achieved, I would avoid any test – which is not required.

» » As I wrote in the article:

In 20132 members of the EMA’s Pharmacokinetics Working Party confirmed unanimously that a test is not expected by the agencies. ‘Demonstration of steady state’ should not be understood in the statistical sense. In­stead, individual pre-dose concentrations should be reported together with spaghetti plots and plots of geometric means. In the discussion it became clear that common sense should prevail.


» I don't have access to the Open Forum Discussion, I tried to find reference no. 02 but I only found session 3 https://custom.cvent.com/6E33084F71EB4DD6B892656072736BFB/files/07323e6d219f47cdae68e89f1f508b6b.pdf, which isn't related to steady state achievement. Can you please provide the steady state related session to support protocol deviation/ amendment?

[image]Hey, you discovered Alfredo’s presentation on the web. ;-)

Just checked all others (links in this post): nothing. However, it was a response by the panelists (all members of the PKWP) to a question of an attendee.

I know that other members of the forum attended this meeting as well. If you don’t believe me, maybe they can join the conversation and confirm that.

See also the end of section 1 in this post which I wrote three days after the meeting when my memory was fresh.

» Please see the article and graph

THX. Where does the graph come from?

P.S.: Please don’t format your entire text in green.

Dif-tor heh smusma 🖖 [image]
Helmut Schütz
[image]

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

2022-04-07 10:58
(83 d 08:06 ago)

@ Helmut
Posting: # 22914
Views: 620
 

 Steady-state achieve­ment

Hi Helmut

» Just checked all others (links in this post): nothing. However, it was a response by the panelists (all members of the PKWP) to a question of an attendee.

Ok, Thank you.

» I know that other members of the forum attended this meeting as well. If you don’t believe me, maybe they can join the conversation and confirm that.

I didn't mean that.

» See also the end of section 1 in this post which I wrote three days after the meeting when my memory was fresh.

Thanks a lot.

» THX. Where does the graph come from?

I don't know, I received it by email as a separate graph. I tried to find the reference but I didn't get it.

» P.S.: Please don’t format your entire text in green.

Ok

One additional question please: spaghetti plots and plots of geometric means are the XY plots that are usually generated in WinNonlin, right? As I'm not familiar with the term "spaghetti plots".

Thank you.
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2022-04-07 15:23
(83 d 03:42 ago)

@ Shatha
Posting: # 22915
Views: 645
 

 “Spaghetti plots”

Hi Shatha,

» » Where does the graph come from?
»
» I don't know, I received it by email as a separate graph. I tried to find the reference but I didn't get it.

Maybe ask the one who sent it to you for the source.

» spaghetti plots and plots of geometric means are the XY plots that are usually generated in WinNonlin, right? As I'm not familiar with the term "spaghetti plots".

“Spaghetti plots” (i.e., overlays of subjects’ time course of concentrations grouped by treatment) is parlance used by the PK community. You plot the lines and omit the data points. Few references.1,2


  1. Bonate PL. Pharmacokinetic-Pharmacodynamic Modeling and Simulation. New York: Springer; 2nd ed. 2011. p. 359.
  2. Gabrielsson J, Weiner D. Pharmacokinetic and Pharmacodynamic Data Analysis: Concepts and Appli­ca­tions. Stock­holm: Apotekar­so­ci­eteten; 5th ed. 2016. p. 335–6, p. 704.

Dif-tor heh smusma 🖖 [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
22,181 posts in 4,647 threads, 1,572 registered users;
online 4 (0 registered, 4 guests [including 1 identified bots]).
Forum time: Wednesday 19:05 CEST (Europe/Vienna)

Nerds don’t just happen to dress informally.
They do it too consistently.
Consciously or not, they dress informally
as a prophylactic measure against stupidity.    Paul Graham

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