Astea
★★  

Russia,
2014-12-28 13:40
(3378 d 02:43 ago)

Posting: # 14181
Views: 11,229
 

 enterohepatic recycling, how to calculate Kel? [PK / PD]

Dear all!
Can you please give me some links where to find the information of how to calculate Kel for drugs with enterohepatic recycling? What part of the curve we should use for analysis? :confused:

Beforehand grateful


Edit: Category changed. [Helmut]

"Being in minority, even a minority of one, did not make you mad"
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-12-28 14:56
(3378 d 01:27 ago)

@ Astea
Posting: # 14182
Views: 11,076
 

 Kel or λz?

Hi Astea,

❝ Can you please give me some links where to find the information of how to calculate Kel for drugs with enterohepatic recycling?


If you are dealing with an oral dose it is practically impossible to claim enterohepatic recycling. A second peak may also be the result of transfer for the stomach to the duodenum, a food effect (increased hepatic blood flow), etc. Only if you see a second peak after an iv dose you have strong hints. Many models exist, and without an IV dose and sampling faeces PK parameters are not identifiable. No idea about the web. [image] it. Standard PK textbooks contain examples.

❝ What part of the curve we should use for analysis? :confused:


The latest one. If you are talking about BA/BE we are only applying NCA (i.e., we are not dealing with PK models: \(\small{\lambda_\textrm{z}\neq k_\textrm{el}}\)). Only make sure that you have at least three samples in the log/linear decline and AUCt covers ≥80% 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
Astea
★★  

Russia,
2014-12-28 15:42
(3378 d 00:41 ago)

@ Helmut
Posting: # 14183
Views: 10,092
 

 Kel or λz?

Thanks a lot for your answer!
I shall explain: we deal with the standard BE study: mycophenolic acid orally. The achieved curves have multi-peak form: there is a second small peak, that could be reffered as enterohepatic recycling (known in literature for MPA). So the question is how to determine the half life of such drug? Do we need to use the second terminal part of the curve?
And additional silly question: three samples in the log/linear decline can include Cmax-sample or there should be three samples after Tmax? And if the latest of them is zero is the regression line correct?


Edit: Full quote removed. Please delete everything from the text of the original poster which is not necessary in understanding your answer; see also this post! [Helmut]

"Being in minority, even a minority of one, did not make you mad"
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-12-28 16:34
(3377 d 23:49 ago)

@ Astea
Posting: # 14184
Views: 10,292
 

 Kel or λz?

Hi Astea,

❝ […] we deal with the standard BE study: mycophenolic acid orally.


Difficult drug to show BE.

❝ The achieved curves have multi-peak form: there is a second small peak, that could be reffered as enterohepatic recycling (known in literature for MPA).


Correct.

❝ So the question is how to determine the half life of such drug? Do we need to use the second terminal part of the curve?


Yes. It might well be possible that you see in some patients no second peak, the second one smaller than the first one or vice versa. This behavior may even change within the same patient. For a nice PK model see Prémaud et al. (2005).*

❝ […] three samples in the log/linear decline can include Cmax-sample…


Not a good idea. Absorption is not (practically) completed at tmax.

❝ … there should be three samples after Tmax?


Yes. If you have a second peak, after thistmax”.

❝ And if the latest of them is zero is the regression line correct?


There is no zero concentration post dose (well, there is but after a very long, long time). I guess you mean below the limit of quantification (BQL)? Even if you believe in zero, you can’t use it. ln(0) is not defined.

For MFA it is possible that you fail to get a reliable estimate of \(\small{\lambda_\textrm{z}}\) in a substantial number of pro­files. Therefore, you have no formal proof that AUCt-∞ < 20% of AUC. Discuss it in the study report (according to the GLs don’t exclude subjects). Since the variability of Cmax is generally higher than the one of AUCt, you can try a sensitivity analysis: BE of Cmax (all subjects) and BE of AUCt (only ones with a reliable estimate).


  • Prémaud A, Debord J, Rousseau A, Le Meur Y, Toupance O, Lebranchu Y, Hoizey G, Le Guellec C, Marquet P. A Double Absorption-Phase Model Adequately Describes Mycophenolic Acid Plasma Profiles in De Novo Renal Transplant Recipients Given Oral Mycophenolate Mofetil. Clin Pharmacokinet. 2005;44(8):837–47. doi 10.2165/00003088-200544080-00005.

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
Astea
★★  

Russia,
2014-12-28 18:48
(3377 d 21:35 ago)

@ Helmut
Posting: # 14187
Views: 10,964
 

 Kel or λz?

I'm very grateful for your answer! Now everything is in the right place.

❝ Difficult drug to show BE.


That's it! We've got CI for Cmax of about 115-140% that is far from the required. On the other hand CI for AUC0t lies fully in the proper limits. It is our first study with failed BE and I anticipate the sponsor's anger. I have no ideas of what to do: the standardized differences don't afford to exclude subjects.
Another serious problem is inadequate design: we have only 4-5 non-zero points for each subject; I mean that the concentration rapidly rises from zero to Cmax and then goes down, but the majority of the samples that we measure are far from this range (=are BQL). So there are no three nonzero samples after second Tmax... For our case "after a very long, long time" is just about half an hour! (We've got curves similarily to those pointed in the article Upadhyaya V., et al., Determination of mycophenolic acid in human plasma by ultra performance liquid chromatography tandem mass spectrometry, Journal of Pharmaceutical Analysis 2014; 4(3):205-216)

❝ Yes. It might well be possible that you see in some patients no second peak, the second one smaller than the first one or vice versa. This behavior may even change within the same patient. For a nice PK model see Prémaud et al. (2005).*


All the patients have second peak in the profile of the curve and this peak is significally smaller than the first.
Thank you for the link but full text of the article is not available.
But you claim that PK model is not necessary for standard BE study, am I right?
Are there any official documents (like FDA recommendations or something else) where I can find the statement that half life calculation is incorrect and isn't neccessary in the case when we have lack of information about the terminal phase of the curve?

"Being in minority, even a minority of one, did not make you mad"
mittyri
★★  

Russia,
2014-12-28 21:25
(3377 d 18:58 ago)

@ Astea
Posting: # 14188
Views: 37,186
 

 Kel or λz?

Dear Astea,

I think this thread should be useful

❝ We've got CI for Cmax of about 115-140% that is far from the required.


It would be a main problem. The λz calculation doesn't make any sense in this case.

❝ We've got curves similarily to those pointed in the article Upadhyaya V., et al., Determination of mycophenolic acid in human plasma by ultra performance liquid chromatography tandem mass spectrometry, Journal of Pharmaceutical Analysis 2014; 4(3):205-216)


The mean plasma profiles look very well in the mentioned article. May be I didn't understand something...

❝ ❝ Yes. It might well be possible that you see in some patients no second peak, the second one smaller than the first one or vice versa. This behavior may even change within the same patient. For a nice PK model see Prémaud et al. (2005).*


❝ All the patients have second peak in the profile of the curve and this peak is significally smaller than the first.

❝ Thank you for the link but full text of the article is not available.


I can help you with this article, please contact me. By the way no new information is presented there

❝ Are there any official documents (like FDA recommendations or something else) where I can find the statement that half life calculation is incorrect and isn't neccessary in the case when we have lack of information about the terminal phase of the curve?


Half life isn't a primary metric, I doubt there's one.
The good solution is to provide two parts of AUC analysis: with and without estimable λz (as Helmut wrote).

Kind regards,
Mittyri
Astea
★★  

Russia,
2014-12-28 22:21
(3377 d 18:02 ago)

@ mittyri
Posting: # 14189
Views: 9,993
 

 Kel or λz?

Thanks for your answer! Yes, it is very helpful.

❝ The mean plasma profiles look very well in the mentioned article. May be I didn't understand something...


I pointed this article only for the reason to show the whole shape of the curve and its features. Their analyses and blood collection points are very good. Unfortunately for us this article was printed later than our study was planned. We have similar mean curves but the data is lacked: when they have 5 points we get only one...

"Being in minority, even a minority of one, did not make you mad"
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-12-29 03:58
(3377 d 12:24 ago)

@ Astea
Posting: # 14192
Views: 9,984
 

 Kel or λz?

Hi Astea,

❝ It is our first study with failed BE …


Sorry to say but I will not be the last one.

❝ … and I anticipate the sponsor's anger.


Tell the sponsor that 10–20% of properly designed and performed studies fail by pure chance. That’s part of the game.

❝ I have no ideas of what to do: the standardized differences don't afford to exclude subjects.


Sure. I suggested just an additional evaluation. Of course in Russia it is difficult. According to the old guideline you have to switch from the primary metric AUCt to AUC if 80% are not covered. The “Red Book” follows the EMA’ GL.

❝ Another serious problem is inadequate design: we have only 4-5 non-zero points for each subject; […]



That’s really bad. I guess your LLOQ was way too high…

❝ For our case "after a very long, long time" is just about half an hour!


I was talking about true zero, not BLQ. ;-)

❝ (We've got curves similarily to those pointed in the article Upadhyaya V., et al.


That’s a strange paper. In Fig. 6 they have not a time-scale, but clicked the wrong button in Excel. That’s a line-plot. Bizarre. The T/R-ratios are close to 1 and the reported CVs were 5.17–6.34%. However, I calculate 13.1–18.4%.
They succeeded in selling the sponsor an extremely overpowered BE-study in 72 subjects with 5,328 plasma samples. Are Indian IECs sleeping or waving all studies through – completely misunderstanding what their job is?
“Furthermore, there was no adverse event during the course of the study.” Not even a single case of headache? Gimme a break.

❝ All the patients have second peak in the profile of the curve and this peak is significally smaller than the first.


Good to know that the second one was smaller.

❝ Thank you for the link but full text of the article is not available.


You’ve got [image].

❝ But you claim that PK model is not necessary for standard BE study, am I right?


PK modeling is not acceptable in BE (only NCA), but it helps to understand what’s going on.

❝ Are there any official documents…


I don’t think so.

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
Astea
★★  

Russia,
2014-12-29 11:38
(3377 d 04:45 ago)

@ Helmut
Posting: # 14198
Views: 9,877
 

 Kel or λz?

Dear Helmut!

❝ Tell the sponsor that 10–20% of properly designed and performed studies fail by pure chance. That’s part of the game.


We are now to prove bio-inequivalence if it is really takes the place. How to calculate power in this case (I use PowerTOST in R)? We have CV, CI for Cmax with the left margin greater then the proper and N (imbalanced study after excluding drop-outs)...

"Being in minority, even a minority of one, did not make you mad"
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2014-12-29 15:14
(3377 d 01:09 ago)

@ Astea
Posting: # 14200
Views: 9,842
 

 TOST

Hi Astea,

❝ We are now to prove bio-inequivalence if it is really takes the place. How to calculate power in this case (I use PowerTOST in R)? We have CV, CI for Cmax with the left margin greater then the proper and N (imbalanced study after excluding drop-outs)...


With a 90% CI of 115–140% bioinequivalence is not formally proven. This would only be the case if the entire CI is outside 80–125%. Note that from the Two One-Sided Tests as proposed by Schuirmann you get two p-values (one for <0.8 and the other one for >1.25). In your case only the one testing for the lower limit will be significant (<0.05). However, since the PE of 127% is already outside the acceptance range you have some strong hints of inequivalence.

The p-values of TOST are not implemented in the current version of PowerTOST. AFAIK Detlew is considering it for one of the next revisions.

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,
2014-12-31 14:41
(3375 d 01:42 ago)

@ Helmut
Posting: # 14207
Views: 9,529
 

 PowerTOST

Hi Astea, hi Helmut,

❝ The p-values of TOST are not implemented in the current version of PowerTOST. AFAIK Detlew is considering it for one of the next revisions.


This next version, V1.2-04, is already on CRAN since 2014-12-19 with the new function pvalues.TOST() contributed by Ben.

Hope this helps.
Happy New Year to All (especially the Power[TOST] users) :party:.

Regards,

Detlew
luvblooms
★★  

India,
2014-12-29 06:51
(3377 d 09:31 ago)

(edited by luvblooms on 2014-12-29 09:19)
@ Helmut
Posting: # 14194
Views: 9,962
 

 Kel or λz?

❝ Hi Helmut


Just want to add something here

❝ If you are dealing with an oral dose it is practically impossible to claim enterohepatic recycling.


There are few examples where hepatic recirculation is very much obivious in oral dose as well. For e.g.Ezetemibe where the amount of ezetimibe recycled into the central compartment was estimated to be approximately 17% to 20% of the total amount absorbed, independent of the volume of distribution

❝ A second peak may also be the result of transfer for the stomach to the duodenum, a food effect (increased hepatic blood flow), etc. Only if you see a second peak after an iv dose you have strong hints.


One more reason sould be differential solubility and absorption across GIT (Diclofenac). The best way to that whether HR is there or not is to use IV dose if no prior published information is available.

~A happy Soul~
nobody
nothing

2014-12-29 22:54
(3376 d 17:28 ago)

@ luvblooms
Posting: # 14204
Views: 9,770
 

 Kel or λz?

eehm I didn'T read the original article (only abstract), but half of the relevant parameters are not identifiable without i.v. data... sorry, but have seen so much nonsense PK evaluations even in the assessment reports published by FDA etc. over the years, don't buy this EHR without additional data.

Kindest regards, nobody
luvblooms
★★  

India,
2014-12-30 06:25
(3376 d 09:58 ago)

@ nobody
Posting: # 14205
Views: 9,689
 

 Kel or λz?

Dear Nobody

❝ eehm I didn'T read the original article (only abstract), but half of the relevant parameters are not identifiable without i.v. data... sorry, but have seen so much nonsense PK evaluations even in the assessment reports published by FDA etc. over the years, don't buy this EHR without additional data.


Ok. One stupid question
"How to determine the absolute bioavailability of drugs as ezetimibe which are virtually insoluble in aqueous media suitable for injection?"

IMHO, in those cases you have to rely on the oral IR tablet/suspension data.

Where extensive glucuronidation in the lumen and liver is involved (leading to formation of glucuronide metabolite), there will be high chances or occurence of hepatic recirculation.

Also, the movement of bile acids in the EHC fluctuates, accelerating with meals and decelerating during fasting. The movement of the EHC is largely determined by the activity of its mechanical pump, the small intestine, whose motility is governed by neurohormonal factors initiated by entrance of food into the small intestine from the stomach. Thus for all practical purposes the “rhythm” of the EHC is determined by the duration of gastric emptying, which in turn is determined by the eating pattern and dietary content. Contraction of the gallbladder is essential for mobilizing the stored bile acids but is not critical for further movement of bile acids, which depends on intestinal propulsion.

Thus if the multiple peak is occurring around the meal time (stimuli for bile secretion, one can see it around 4-6 hr post dose during lunch time, 8 -10 hrs post dose around snacks time and 12-14 hrs post dose around dinner time), it most probably will be because of EHC as food acts as stimuli for bile acid secretion.

[image]

Would be happy if someone can correct me if I got this one wrong.

~A happy Soul~
UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,638 registered users;
91 visitors (0 registered, 91 guests [including 12 identified bots]).
Forum time: 16:23 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