Charl
●    

2007-07-23 12:44
(6093 d 05:01 ago)

Posting: # 924
Views: 12,193
 

 PK Vs Analytical [Bioanalytics]

Hi :-)

what is the best way to decide on the LLOQ depending on pharmacokinetics point of view, rather than depending on the analytical parameters to decide?

Is there a proper way to estimate the linearity range from PK data rather than analytical approach?

My point can we work on the analytical method development depending on the PK and statistical output to design our analytical method? :-|
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2007-07-23 20:04
(6092 d 21:42 ago)

@ Charl
Posting: # 926
Views: 10,978
 

 PK Vs Analytical

Dear Charl!

❝ what is the best way to decide on the LLOQ depending on pharmacokinetics point of view, rather than depending on the analytical parameters to decide?


That’s the only way to do it (we had this already)!

❝ Is there a proper way to estimate the linearity range from PK data rather than analytical approach?


Analytics is just a tool serving PK. You must cover the entire range from the highest Cmax observed in any of the subjects to a suitable LLOQ (covering 80% of AUC in the subject with the lowest levels).
The upper range is not so critical, because you may dilute samples above the ULOQ and re-analyse them. Of course the dilution step must be validated. ;-)
In BE linearity generally is not that critical, but in PK (e.g., multi-phasic models after an i.v. administration) sometimes it may be rather tough due to limited detector linearity in MS and fluorescence.

❝ My point can we work on the analytical method development depending on the PK and statistical output to design our analytical method? :-|


Sure, the PK-people should tell you what concentrations they expect in the study.
They may refer to previous studies, literature, or PK-simulations…

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
Charl
●    

2007-07-24 12:31
(6092 d 05:14 ago)

@ Helmut
Posting: # 929
Views: 10,760
 

 PK Vs Analytical

❝ Sure, the PK-people should tell you what concentrations they expect in the study.

❝ They may refer to previous studies, literature, or PK-simulations...


Am looking for a more specific way to determine linearity, from PK point of view,..... :surprised:
If we combined data in a "mathmatical equation" such as Absorption, excretion, inter-subject variabilty, Dose, rate of release in body......
Can we finally have a Model that would give us a close estimation to what we usually end up by the end of the study.
How far can we predict to what degree of details can we go.... :ponder:


Edit: Please have a look at the policy (capitalization). [Jaime]
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2007-07-24 15:42
(6092 d 02:03 ago)

@ Charl
Posting: # 931
Views: 11,138
 

 PK Vs Analytical

Dear Charl!

❝ Am looking for a more specific way to determine linearity, from PK point of view,.....


I’m still not getting your point – maybe since your are posting in the ‘Bioanalytics’-category.
You should define a calibration function (which must not be linear, but describing a continous function of detector response from known concentrations) within LLOQ and ULOQ.
LLOQ and ULOQ are defined by PK requirements, whereas the calibration function depends on the analytical method.

❝ If we combined data in a "mathmatical equation" such as Absorption, excretion, inter-subject variabilty, Dose, rate of release in body......

Can we finally have a Model that would give us a close estimation to what we usually end up by the end of the study.

How far can we predict to what degree of details can we go....


That’s an endless story – for sure nothing to be explained in a couple of posts (many people devote their entire lifetimes on small fractions of these topics). I’ll give you some keywords:
  • Biopharmaceutics Classification System (BCS)
  • Biorelevant dissolution media
  • In vivoin vitro correlation (IVIVC)
  • Population Pharmacokinetics (PopPK)
Yes, there’s software outside; you may e.g., go with Globomax’ NONMEM/PDx-IVIVC, Pharsight’s WinNonlin/IVIVC Toolkit, or even simulations plus’ GastroPlus solving 80 simultaneous differential equations for $50,000…

Anyhow, if you are in doubt whether your method is suitable, why don’t you opt for a small pilot study?
http://www.fda.gov/cder/guidance/5356fnl.pdf FDA states at III.A.2.:

If the sponsor chooses, a pilot study in a small number of subjects can be carried out before proceeding with a full BE study. The study can be used to validate analytical methodology, assess variability, optimize sample collection time intervals, and provide other information.


Edit: FDA-link corrected to latest archived copy. [Helmut]

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
Charl
●    

2007-07-24 17:02
(6092 d 00:43 ago)

@ Helmut
Posting: # 932
Views: 10,874
 

 PK Vs Analytical

Dear HS... :-)

thanks, your advice was helpfull.
drshiv
★    

India,
2007-07-31 10:47
(6085 d 06:58 ago)

@ Charl
Posting: # 946
Views: 10,793
 

 PK Vs Analytical

Dear Charl,

I hope the following explanation helps you.

Determination of Lower Limit of Quantitation (LLOQ)
LLOQ should be sufficiently low to allow characterization of 4-5 T1/2
AUC0-t/AUCinf ration should be >80% (ie., residual area <20%)
LLOQ should allow proper characterization of baseline levels of endogenous substance (unless they are insignificant, eg., less than 5% of Cmax)
LLOQ values should be less than 5% of observed Cmax
Examples of LLOQ determination methods:
Expected mean Cmax divided by 32 (represents 5 T1/2 assuming a one-compartment model)
Should also be based on the PK profile

Dr. Shiv
Charl
●    

2007-07-31 16:07
(6085 d 01:38 ago)

@ drshiv
Posting: # 947
Views: 10,640
 

 PK Vs Analytical

❝ Should also be based on the PK profile


Dear Dr. Shiv

❝ Determination of Lower Limit of Quantitation (LLOQ)

❝ LLOQ should be sufficiently low to allow characterization of 4-5 T1/2


this is general, instrument sensitivity might not allow us to go to the disered LLOQ if we are chasing only a 4-5 T1/2.

❝ LLOQ should allow proper characterization of baseline levels of endogenous substance (unless they are insignificant, eg., less than 5% of Cmax)

❝ LLOQ values should be less than 5% of observed Cmax


in what since you are conecting LLOQ to be 5% les than Cmax, I conclude from your words that my entire linearty range lies in the 5% region??? :-|

❝ Examples of LLOQ determination methods:

❝ Expected mean Cmax divided by 32 (represents 5 T1/2 assuming a one-compartment model)


what if am working on non-compartmental model does this still apply.... ;-)
drshiv
★    

India,
2007-08-07 18:37
(6077 d 23:08 ago)

@ Charl
Posting: # 967
Views: 10,732
 

 PK Vs Analytical

Dear Charl,

<5% assumption is to again monitor concentration upto 5 half-lives.

All the assumptions are model independent. Again, in BE studies only non-compartmental models are used. It applies to that also.

Dr. Shiv


Edit: Full quote removed. Please see this post! [HS]
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2007-08-11 22:09
(6073 d 19:37 ago)

@ drshiv
Posting: # 972
Views: 10,822
 

 NCA 'model independent'?

Dear Dr. Shiv!

❝ <5% assumption is to again monitor concentration upto 5 half-lives.

❝ All the assumptions are model independent. Again, in BE studies only non-compartmental models are used.


I certainly agree with you that all guidelines recommend NCA for BE.
But we must not forget that NCA is not really independent from a PK model.
When you are talking about half-life, you are implying (at least for the ‘terminal’ phase within the profile) first order elimination.
The ‘5% rule’ is doubtful for a Michaelis-Menten type of elimination, and simply wrong for a zero-order process (e.g., ethanol).
So again, study planing in all its aspects (number and distribution of sampling points, length of sampling interval,…) must always be guided by a thorough knowledge of the drug’s PK.

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
drshiv
★    

India,
2007-08-12 15:08
(6073 d 02:37 ago)

@ Helmut
Posting: # 977
Views: 10,698
 

 NCA 'model independent'?

Dear HS,

I am really on a high on this subject. Thanks we are really discussing technical aspects of BE not just guidelines (which I am bored of)!!!.
I really welcome your comments.

My answer was keeping in mind only BE studies. Nearly all most all the drugs undergo first-order elimination process. Under rare exception, even, drug delivery systems designed for zero order release shows first order elimination. I did not take into consideration when answering 5% rule, the zero-order processess and Michaelis-Menten type elimination. NCA is useful tool for typical single dose studies. At lower dose even drugs undergoing Michaelis-Menten type drugs show first order elimination profiles. Only at higher doses they show the dome shaped profile which is indicative of saturation. I agree with you.

Dr. Shiv


Edit: Full quote removed. Please see this post! [HS]
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2007-08-12 20:51
(6072 d 20:54 ago)

@ drshiv
Posting: # 978
Views: 11,372
 

 Numerical integration

Dear Dr. Shiv!

❝ I am really on a high on this subject. Thanks we are really discussing technical aspects of BE not just guidelines (ehich I am bored of)!!!.


Me too! :ok:

❝ NCA is useful tool for typical single dose studies…


Agree. If you are interested in a more thorough discussion:
Everybody should be aware that the linear trapezoidal rule is close to the worst numerical integration algorithm (just after the rectangular rule).
There are other algorithms available (log-trapezoidal, lin/log-trapezoidal, Lagrange, Splines, Simpson’s rule, Runge-Kutta). This was pointed out almost 30 years ago by Yeh and Kwan – without further notice by the scientific community.
At least modern software allows selection of the log-trapezoidal rule, or even better lin-up/log-down. Funny enough in all packages I know the linear trapezoidal rule is set as the default leaving the unwary investigator with a suboptimal method. :-D

  1. Yeh, KC and KC Kwan
    A Comparison of Numerical Integrating Algorithms by Trapezoidal, Lagrange, and Spline Approximation
    J Pharmacokin Biopharm 6(1), 79–98 (1978)
  2. RD Purves
    Optimum Numerical Integration Methods fo Estimation of Area-Under-the-Curve (AUC) and Area-under-the-Moment-Curve (AUMC)
    J Pharmacokin Biopharm 20(3), 211–26 (1992)

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
drshiv
★    

India,
2007-08-14 10:29
(6071 d 07:16 ago)

@ Helmut
Posting: # 981
Views: 10,659
 

 Numerical integration

Hi HS,

Yes, you are right, trapezoidal rule is just an approximation method. I have read the paper of Yeh and Kwan way back in 1989 and used it to write programs for my PhD research. I have also used 5 part Runge-Kutta method for area calculation and also solving numerical ordinary differential equations. You have rightly put that people must use different methods to know the significance of these methods in the area of pharmacokinetics for obtaining numerical solutions.

Thanks for bringing these issues for open discussion.

Dr. Shiv


Edit: Full quote removed. Please see this post! [HS]
UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,638 registered users;
66 visitors (1 registered, 65 guests [including 4 identified bots]).
Forum time: 16:46 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