Time-varying PK DURING the course of study

Dear nmusers, I found it is difficult to account for age varying in PK parameter during the course of the study. Is there any way on how to define icreasing age as a covariate in the NONMEM data file? A short example would be very useful for me. thanks for your help. khaled

Dear Khaled, Time varying covariates are be included in a number of ways. This paper by Ulrika Wählby et al. explores this subject in some depth. Wählby, U., Thomson, A. H., Milligan, P. A., & Karlsson, M. O. (2004). Models for time-varying covariates in population pharmacokinetic-pharmacodynamic analysis. British journal of clinical pharmacology, 58(4), 367-377. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15373929 Kind regards Dinko Rekic

Hi Khaled, You can simply include age as a new column. The complexity is that the solution will change at the time that the row happens. So, if the change in age between rows in the table is small relative to the total covariate effect (like age changes from 20 to 21 which makes a 1% change in clearance) then you can likely just use the new column. Also, if you are using an algebraic model (instead of compartmental), then you can just put the new age in place. If the change in age between rows is large relative to the total covariate effect (like age changes from 20 to 50 which makes a 50% change in clearance), then you will need to do additional coding so that the age varies continuously. You will need to add next age (AGEN) and next time (TIMN) columns; that would just be taking the age and time columns and put them in the next row down for the same subject (making sure that you don't do a data management error of shifting them between subjects). Then you will need to use differential equations if you have a compartmental model, and you will need to add an imputed age-just a linear model of age like. ; imputed age AGEI = (AGE - AGEN)/(TIME - TIMN)*(T-TIME) + AGE Then use AGEI as your age in the differential equations. Thanks, Bill