Simultaneous vs sequential for modeling parent AND metabolites in pop PK

3 messages 2 people Latest: Dec 10, 2008

RE: Simultaneous vs sequential for modeling parent AND metabolites in pop PK

From: Xiaofeng Wang Date: December 09, 2008 technical
The appropriate approach is to fit parent compound and the metabolites simultaneously, since it can help uniquely define the PK parameters, especially the CL. As we know that CL is a lumped parameter without metabolite information. Sequential estimation is the approach when there is no better way to solve the problem in simultaneous fitting due to numerical problem. This is different situation from PK-PD fitting. In the situation if PD has impact on PK, also, simultaneous fitting is the way to go. xiaofeng Xiaofeng Wang, PhD Oncology, Novartis (862)778-8856 (o) "Xiao, Alan" <Alan.Xiao Sent by: owner-nmusers 12/09/2008 01:30 PM To "Bachman, William" <William.Bachman cc Subject RE: [NMusers] Simultaneous vs sequential for modeling parent AND metabolites in pop PK Dear All, Thanks for your response and I'm sorry for the confusion. I'm talking about the sequential/simultaneous modeling to fit parent concentrations AND metabolites in pop PK, not about PD data at all. That is for sequential approach, you develop a model to fit the parent data first and then fix the PK parameters for parents to develop a model to fit the metabolite. While, for simultaneous approach, you develop a model to fit both parent and metabolites simultaneously (to simultaneously estimate parameters for both parent and metabolites). Alan
Quoted reply history
-----Original Message----- From: Bachman, William [mailto:William.Bachman Sent: Tuesday, December 09, 2008 11:26 AM To: Xiao, Alan Cc: nmusers Subject: RE: [NMusers] Simultaneous vs sequential for modeling parent AND metabolites in pop PK The argument against the simultaneous approach is that the PD data can "drive" the PK model, particulary since the PD data usually has more variability. -----Original Message----- From: owner-nmusers On Behalf Of Xiao, Alan Sent: Tuesday, December 09, 2008 11:02 AM To: nmusers Subject: [NMusers] Simultaneous vs sequential for modeling parent AND metabolites in pop PK Dear All, I know this is an old topic but would like to see the statistics. When you have to develop a pop PK model for both parent and active metabolites, which approach do you prefer or have you used most: simultaneous or sequential? Which way do you think is more scientific? I heard comments saying that the simultaneous approach is not scientific. Thanks, Alan

Re: Simultaneous vs sequential for modeling parent AND metabolites in pop PK

From: Nick Holford Date: December 09, 2008 technical
Hi, Experimental studies (rather than just opinion) can be found here: 1. Zhang L, Beal SL, Sheiner LB. Simultaneous vs. sequential analysis for population PK/PD data I: best-case performance. J Pharmacokinet Pharmacodyn. 2003;30(6):387-404. 2. Zhang L, Beal SL, Sheiner LB. Simultaneous vs. sequential analysis for population PK/PD data II: robustness of methods. J Pharmacokinet Pharmacodyn. 2003;30(6):405-16. 3. Proost JH, Schiere S, Eleveld DJ, Wierda JM. Simultaneous versus sequential pharmacokinetic-pharmacodynamic population analysis using an iterative two-stage Bayesian technique. Biopharm Drug Dispos. 2007;28(8):455-73. My interpretation of these studies is that one should use a sequential approach to build the model then try a simultaneous fit. If the PK part of the model changes 'importantly' (subjective decision) with the simultaneous fit then this can be a clue to model misspecification in the link between the PK and PD parts of the model (see second paper by Zhang et al). Nick Hussein, Ziad wrote: > Hi Alan, > > I just had a very recent experience few weeks ago for a sequential PopPK > for parent and metabolite that was submitted to the FDA and they came > back and asked for simultaneous modelling. > > Whether this is scientific or not the FDA view should be taken into > consideration. > > Kind regards, > Ziad >
Quoted reply history
> -----Original Message----- > From: owner-nmusers > On Behalf Of Xiao, Alan > Sent: 09 December 2008 16:02 > To: nmusers > Subject: [NMusers] Simultaneous vs sequential for modeling parent AND > metabolites in pop PK > > Dear All, > > I know this is an old topic but would like to see the statistics. > > When you have to develop a pop PK model for both parent and active > metabolites, which approach do you prefer or have you used most: > simultaneous or sequential? Which way do you think is more scientific? I > heard comments saying that the simultaneous approach is not scientific. > > Thanks, > > Alan >

Re: Simultaneous vs sequential for modeling parent AND metabolites in pop PK

From: Nick Holford Date: December 10, 2008 technical
James, All is well with me and it seems things continue to be (w)right for you :-) Non-proportionality (aka non-linearity) of metabolite formation from parent is not really an issue. With the right design (i.e. suitable doses of parent) then this can be discovered from just giving the parent. But to really know how much of the parent is eventually transformed to the metabolite needs additional information i.e. direct administration of the metabolite. Best wishes, Nick James G Wright wrote: > Hi Nick, > > I hope all is well with you - good to see you are keeping the nmusers in > line :-). > > I am not sure I agree with your second paragraph as I have understood > it. When I fit a parent-metabolite model, I estimate CL/F and V/F for > the metabolite. Frequently, I hear the widespread misconception that I > have assumed all of the parent goes to metabolite, but this would only > be true if I claimed to have estimated CL, rather than CL/F. This is > exactly the same as if I analyse the parent after oral administration > (without IV) - we don't assume all the drug is absorbed, we simply > estimated the ratio of CL and V to F. > > The real assumption lies in the form of the link between parent and > metabolite - for example, that it is linearly formed from parent in > plasma. It is this assumption that may need to be more rigorously > evaluated, and exactly the point highlighted in your first paragraph. > For example, there can be an apparent delay in metabolite formation > relative to parent plasma concentrations and/or the metabolite may be > formed during the first-pass. > > Best regards, James > > PS One can get technical and claim there needs to be a correction for > molecular weight in F for the metabolite, but the importance of this > depends on how the parameter will actually be used. > > James G Wright PhD > Scientist > Wright Dose Ltd > Tel: 44 (0) 772 5636914 > >
Quoted reply history
> -----Original Message----- > From: owner-nmusers > On Behalf Of Nick Holford > Sent: 09 December 2008 20:52 > To: nmusers > Subject: Re: [NMusers] Simultaneous vs sequential for modeling parent > AND metabolites in pop PK > > Alan, > > The comments about sequential vs simultaneous modelling apply for any > kind of multivariate approach. The 'driver' model e.g. parent conc for > metabolite or the 'driven' model e.g. metabolite from parent will be > dependent on having a good driver model first. If the driver plus driven > > dont do well together with a simultaneous fit then this is a clue to > model misspecification. > > Parent-metabolite models nearly always have to make at least one > unverifiable assumption if the metabolite is not given directly. (e.g. > one may assume all of the parent goes to metabolite OR assume a volume > for the metabolite). PKPD models also have unverifiable assumptions e.g. > > concentration at the site of the drug effect. > > Nick > > Xiao, Alan wrote: > >> Dear All, >> >> Thanks for your response and I'm sorry for the confusion. >> >> I'm talking about the sequential/simultaneous modeling to fit parent >> concentrations AND metabolites in pop PK, not about PD data at all. >> >> That is for sequential approach, you develop a model to fit the parent >> > data > >> first and then fix the PK parameters for parents to develop a model >> to fit the metabolite. While, for simultaneous approach, you develop a >> > model to > >> fit both parent and metabolites simultaneously >> (to simultaneously estimate parameters for both parent and >> > metabolites). > >> Alan >> >> -----Original Message----- >> From: Bachman, William [mailto:William.Bachman >> Sent: Tuesday, December 09, 2008 11:26 AM >> To: Xiao, Alan >> Cc: nmusers >> Subject: RE: [NMusers] Simultaneous vs sequential for modeling parent >> AND metabolites in pop PK >> >> >> The argument against the simultaneous approach is that the PD data can >> "drive" the PK model, particulary since the PD data usually has more >> variability. >> >> -----Original Message----- >> From: owner-nmusers >> > [mailto:owner-nmusers > >> On Behalf Of Xiao, Alan >> Sent: Tuesday, December 09, 2008 11:02 AM >> To: nmusers >> Subject: [NMusers] Simultaneous vs sequential for modeling parent AND >> metabolites in pop PK >> >> Dear All, >> >> I know this is an old topic but would like to see the statistics. >> >> When you have to develop a pop PK model for both parent and active >> metabolites, which approach do you prefer or have you used most: >> simultaneous or sequential? Which way do you think is more scientific? >> > I > >> heard comments saying that the simultaneous approach is not >> > scientific. > >> Thanks, >> >> Alan >>