Dear sirs, I’m trying to use PKPD. I use PKPD model with Ke0 link with winnonlin PK model 13 and PD model 106. However, the processing halted due to floating point error ( ZERO DIVIDE). I input PK parameter values and dose but not stripping dose since apparently WNL did not ask it. I think that this may be the problem. Can you help me? Thanks Roberto [file name=do1pkpd.phxproj size=423863]Certara | Drug Development Solutions (414 KB)
Hi Roberto, I took a quick look at this and I would propose using a micro constant model, in fact number 12 i.e. 2 com with a lag instead initially and in your PKPD model. (assuming you dose is indeed Extravascular this gives a slightly better fit on the Cmax However there’s still work to do because combined with the PD 106 (Sigmoid Emax with baseline) you still aren’t collapsing the hysteresis. take a look at the attached project and you’ll see I tried a few models and ranked them by SBC to decide the PK fit. Also note I linked the PK parameters straight across to the link model Lastly I took a quick look at building up the same model in a Phoenix model but as a simultaneous fit just so you can see how you might do that. 
Please take some time to look at this (I’ve not really validated it anyway yet). In the mean time I’ll try and find out any tips re using macro constant models in link models Simon [file name=do1pkpd_micro.phxproj size=1720938]Certara | Drug Development Solutions (1.64 MB)
