Mu referencing in covariate model

Dear Group, I want to test CRCL as a covariate on clearance as follows: TVCL = THETA(1)*(CLCR/120)**THETA(2). I am using EM methods for estimation and the following code can be used as per the manual: MU_1= LOG(THETA(1))+THETA(6)*LOG(CLCR/120) CL=EXP(MU_1+ETA(1)) The issue with this code for my dataset is that CLCR is not constant within individual with different observations and violates the mu-referencing rules. I appreciate your suggestions and help on this. Regards, Ayyappa