AW: time dependant function

Dear Sebastien, thanks for your reply. I never heard about DDEs before. There is also no workaround for NONMEM? I would like to use this model for parameter estimation and not for simulation only. Kind regards, Hauke Von: [email protected] [mailto:[email protected]] Im Auftrag von Sebastien Bihorel Gesendet: Freitag, 25. November 2011 12:14 An: Hauke Rühs Cc: [email protected] Betreff: Re: [NMusers] time dependant function Dear Hauke, If I am not mistaken, the model that you are trying to reproduce is a lifespan indirect response (LIDR) model. The direct implementation of such models requires a solver of delay differential equations (DDE), which is not available in NONMEM or Berkeley Madonna. Other software platforms such as Matlab or R provides DDE solvers but require quite a bit of coding if you don't use special packages. If you want to simulate the model and do not need a population estimation approach, I would suggest the scaRabee package which has been designed for PKPD modeling in R and accommodate LIDR models ( http://cran.r-project.org/web/packages/scaRabee/index.html). Dataset preparation and model coding is close to NONMEM conventions. I think that some mathematical and dataset tricks can also allow the implementation LIDR models using the ordinary differential equations solver available in NONMEM, ADAPT, or Berkeley Madonna. However, my understanding is that these methods may only apply to "simple" experimental settings and/or LIDR models. For illustration, see: Perez-Ruixo JJ, Kimko HC, Chow AT, Piotrovsky V, Krzyzanski W, Jusko WJ. Population cell life span models for effects of drugs following indirect mechanisms of action. J Pharmacokinet Pharmacodyn. 2005 Dec;32(5-6):767-93. Pérez-Ruixo JJ, Krzyzanski W, Bouman-Thio E, et al. Pharmacokinetics and pharmacodynamics of the erythropoietin Mimetibody construct CNTO 528 in healthy subjects. Clin Pharmacokinet. 2009;48(9):601-13. Woo's PhD thesis at SUNY at Buffalo. Hope that helps Sebastien Hauke Rühs wrote: Dear NMusers, I was trying to build a published model (Woo et al, Cancer Chemother. Pharmacol, 2008) in NONMEM and Berkeley Madonna, where the differential equations contain time dependant functions. In this particular function the amount in one compartment at a certain time point is addressed. E.g. S(t)=1+(SMAX*EPO(t))/(SC50+EPO(t)) and in the ODEs S(t), S(t-TP1), S(t-TP2) and so on is needed. As EPO(t) is driven by a feedback loop, I didn't succeed by just starting multiple compartments with a delay. Any ideas are welcome. Kind regards, Hauke cid:[email protected]_bonnKlinische Pharmazie Hauke Rühs Apotheker Pharmazeutisches Institut Klinische Pharmazie An der Immenburg 4 53121 Bonn Tel: 0228 73-5781 Fax: 0228 73-9757 http://www.klinische-pharmazie.info/ www.klinische-pharmazie.info

Dear Hauke, Have a look at the first and second reference I listed in my first reply. They should describe several ways or tricks to implement DDEs in NONMEM. These tricks can be useful to you... or not. Be aware that they cannot be applied to all LIDR. Sebastien Hauke Rühs wrote: > Dear Sebastien, > > thanks for your reply. I never heard about DDEs before. There is also no workaround for NONMEM? I would like to use this model for parameter estimation and not for simulation only. > > Kind regards, > > Hauke > > *Von:* [email protected] [ mailto: [email protected] ] *Im Auftrag von *Sebastien Bihorel > > *Gesendet:* Freitag, 25. November 2011 12:14 > *An:* Hauke Rühs > *Cc:* [email protected] > *Betreff:* Re: [NMusers] time dependant function > > Dear Hauke, > > If I am not mistaken, the model that you are trying to reproduce is a lifespan indirect response (LIDR) model. The direct implementation of such models requires a solver of delay differential equations (DDE), which is not available in NONMEM or Berkeley Madonna. Other software platforms such as Matlab or R provides DDE solvers but require quite a bit of coding if you don't use special packages. If you want to simulate the model and do not need a population estimation approach, I would suggest the scaRabee package which has been designed for PKPD modeling in R and accommodate LIDR models ( http://cran.r-project.org/web/packages/scaRabee/index.html ). Dataset preparation and model coding is close to NONMEM conventions. > > I think that some mathematical and dataset tricks can also allow the implementation LIDR models using the ordinary differential equations solver available in NONMEM, ADAPT, or Berkeley Madonna. However, my understanding is that these methods may only apply to "simple" experimental settings and/or LIDR models. For illustration, see: > > Perez-Ruixo JJ, Kimko HC, Chow AT, Piotrovsky V, Krzyzanski W, Jusko WJ. Population cell life span models for effects of drugs following indirect mechanisms of action. J Pharmacokinet Pharmacodyn. 2005 Dec;32(5-6):767-93. Pérez-Ruixo JJ, Krzyzanski W, Bouman-Thio E, et al. Pharmacokinetics and pharmacodynamics of the erythropoietin Mimetibody construct CNTO 528 in healthy subjects. Clin Pharmacokinet. 2009;48(9):601-13. > > Woo's PhD thesis at SUNY at Buffalo. > > Hope that helps > > Sebastien > > Hauke Rühs wrote: > > Dear NMusers, > > I was trying to build a published model (Woo et al, Cancer Chemother. Pharmacol, 2008) in NONMEM and Berkeley Madonna, where the differential equations contain time dependant functions. In this particular function the amount in one compartment at a certain time point is addressed. E.g. S(t)=1+(SMAX*EPO(t))/(SC50+EPO(t)) and in the ODEs S(t), S(t-TP1), S(t-TP2) and so on is needed. As EPO(t) is driven by a feedback loop, I didn't succeed by just starting multiple compartments with a delay. Any ideas are welcome. > > Kind regards, > > Hauke > > cid:[email protected]_bonnKlinische Pharmazie > > Hauke Rühs > > Apotheker > > Pharmazeutisches Institut > > Klinische Pharmazie > > An der Immenburg 4 > > 53121 Bonn > > Tel: 0228 73-5781 > > Fax: 0228 73-9757 > > www.klinische-pharmazie.info http://www.klinische-pharmazie.info/