RE: PK and PD variability

Dear Joseph, I promised to give some feedback on the suggestions you made. Estimating IIV on F indeed results in a rather high shrinkage for this parameter (about 65%). However, I don’t think that this might be an issue in my situation since I will not use EBEs of F for any purpose. The idea was only to separate the variability in ED50 into PD and PK related variation. Actually, I think this approach is somewhat similar to the K-PD modelling explained by Jacqmin et al. 2008. I also applied a simulation-reestimation procedure, that is, simulating the model including IIV on F and reestimating with the original model without F. Only IIV on ED50 was affected, as expected. Joseph, the approach you described also worked well and led to the same conclusion: the interindividual variability in the response is dominated by the variability of the PD and to a much lesser extent by PK. Thank you once again for your suggestions. PS: I would still be interested if anyone had a similar problem and how it was solved. Andreas Lindauer Pharmacokineticist Pharmacokinetics and Metabolism R&D Center. Ferrer Internacional S.A. Juan de Sada 32, 08028 Barcelona Tel +34 93 509 3265 Fax +34 93 411 2764 [email protected] www.ferrergrupo.com ¿Necesita imprimir este mensaje? Protejamos el medio ambiente. Li cal imprimir aquest missatge? Protegim el medi ambient. Do you need to print this message? Let's protect the environment. Este mensaje, y en su caso, cualquier fichero anexo al mismo, puede contener información confidencial, siendo para uso exclusivo del destinatario, quedando prohibida su divulgación, copia o distribución a terceros sin la autorización expresa del remitente. Si Vd. ha recibido este mensaje erróneamente, se ruega lo notifique al remitente y proceda a su borrado. Gracias por su colaboración. This message and its annexed files may contain confidential information which is exclusively for the use of the addressee. It is strictly forbidden to distribute copies to third parties without the explicit permission of the sender. If you receive this message by mistake, please notify it to the sender and make sure to delete it. Thank you for your kind cooperation. "Standing Joseph (GREAT ORMOND STREET HOSPITAL FOR CHILDREN NHS TRUST)" <[email protected]> 11/03/2011 16:46 Para "[email protected]" <[email protected]> cc Asunto RE: [NMusers] PK and PD variability Andreas, I think the method you describe will not give the required variability - what will happen is that during $ESTIMATION NONMEM will assign individual F values that best fit the model, they won't necessarily vary with a CV of 55%. (You can prove this to yourself by outputting F in a table and plotting it/checking if the ETA values really do have a CV of 55%, or looking at shrinkage on that parameter). So I suggest this: STEP 1: Simulate new doses for your dataset with variability $INPUT ...DOSE... ..... NDOSE = DOSE*EXP(ETA(XXX)) .... $OMEGA XXX ; put variance here $SIM ; Just do 1 sample but you might want to try a few runs with different seed numbers if your dataset is small ... $TABLE ....NDOSE NDOSE is now your new dose, that have a mean value of the actual dose given but will be simulated to vary according to $OMEGA STEP 2: Put variable doses into dataset and estimate Emax model i.e. put NDOSE instead of DOSE in your dataset Repeat steps 1 and 2, but change the OMEGA in step 1 to see the effect of decreasing variability in F on parameter estimates in step 2. BW, Joe